The presentation of this document has been augmented to identify changes from a previous version. Three kinds of changes are highlighted: new, added text, changed text, and deleted text.


W3C

XML Path Language (XPath) 4.0 WG Review Draft

W3C Editor's Draft 29 November 2024

This version:
https://qt4cg.org/specifications/xpath-40/
Most recent version of XPath:
https://qt4cg.org/specifications/xpath-40/
Most recent Recommendation of XPath:
https://www.w3.org/TR/2017/REC-xpath-31-20170321/
Editor:
Michael Kay, Saxonica <mike@saxonica.com>

Please check the errata for any errors or issues reported since publication.

See also translations.


Abstract

XPath 4.0 is an expression language that allows the processing of values conforming to the data model defined in [XQuery and XPath Data Model (XDM) 4.0]. The name of the language derives from its most distinctive feature, the path expression, which provides a means of hierarchic addressing of the nodes in an XML tree. As well as modeling the tree structure of XML, the data model also includes atomic items, function items, maps, arrays, and sequences. This version of XPath supports JSON as well as XML, and adds many new functions in [XQuery and XPath Functions and Operators 4.0].

XPath 4.0 is a superset of XPath 3.1. A detailed list of changes made since XPath 3.1 can be found in J Change Log.

Status of this Document

This is a draft prepared by the QT4CG (officially registered in W3C as the XSLT Extensions Community Group). Comments are invited.


1 Introduction

Changes in 4.0 

  1. Use the arrows to browse significant changes since the 3.1 version of this specification.

  2. Sections with significant changes are marked Δ in the table of contents.

XPath was originally designed as an expression language to address the nodes of XML trees; it has since been extended to address a variety of non-XML data sources. XPath gets its name from its use of a path notation for navigating through the hierarchical structure of an XML document; similar capabilities for navigating JSON structures were added in versions 3.0 and 3.1. XPath uses a compact, non-XML syntax, allowing XPath expressions to be embedded within URIs and to be used as XML attribute values. XPath is designed to be embedded in (or invoked from) other host languages, including both languages specialized towards XML processing (such as [XSL Transformations (XSLT) Version 4.0] and XSD), and general purpose programming languages such as Java, C#, Python, and Javascript. The interface between XPath and its host language is formalized with an abstract definition of a static and dynamic context, made available by the host language to the XPath processor.

[Definition: A host language for XPath is any environment that provides capabilities for XPath expressions to be defined and evaluated, and that supplies a static and dynamic context for their evaluation. ]

[Definition: XPath 4.0 operates on the abstract, logical structure of an XML document or JSON object rather than its surface syntax. This logical structure, known as the data model, is defined in [XQuery and XPath Data Model (XDM) 4.0].]

XPath 4.0 is a subset of XQuery 4.0. In general, any expression that is syntactically valid and executes successfully in both XPath 4.0 and XQuery 4.0 will return the same result in both languages. There are a few exceptions to this rule:

Because these languages are so closely related, their grammars and language descriptions are generated from a common source to ensure consistency.

XPath 4.0 also depends on and is closely related to the following specifications:

Note:

The XML-based syntax for XQuery known as XQueryX is no longer maintained.

This document specifies a grammar for XPath 4.0, using the same basic EBNF notation used in [XML 1.0]. Unless otherwise noted (see A.3 Lexical structure), whitespace is not significant in expressions. Grammar productions are introduced together with the features that they describe, and a complete grammar is also presented in the appendix [A XPath 4.0 Grammar]. The appendix is the normative version.

In the grammar productions in this document, named symbols are underlined and literal text is enclosed in double quotes. For example, the following productions describe the syntax of a static function call:

FunctionCall ::= EQName ArgumentList
/* xgc: reserved-function-names */
/* gn: parens */
ArgumentList ::= "(" ((PositionalArguments ("," KeywordArguments)?) | KeywordArguments)? ")"

The productions should be read as follows: A function call consists of an EQName followed by an ArgumentList. The argument list consists of an opening parenthesis, an optional list of one or more arguments (separated by commas), and a closing parenthesis.

This document normatively defines the static and dynamic semantics of XPath 4.0. In this document, examples and material labeled as “Note” are provided for explanatory purposes and are not normative.

2 Basics

2.1 Terminology

Changes in 4.0  

  1. The EBNF operators ++ and ** have been introduced, for more concise representation of sequences using a character such as "," as a separator. The notation is borrowed from Invisible XML.  [Issue 1366 PR 1498]

The basic building block of XPath 4.0 is the expression, which is a string of [Unicode] characters; the version of Unicode to be used is implementation-defined. The language provides several kinds of expressions which may be constructed from keywords, symbols, and operands. In general, the operands of an expression are other expressions. XPath 4.0 allows expressions to be nested with full generality.

Note:

This specification contains no assumptions or requirements regarding the character set encoding of strings of [Unicode] characters.

Like XML, XPath 4.0 is a case-sensitive language. Keywords in XPath 4.0 use lower-case characters and are not reserved—that is, names in XPath 4.0 expressions are allowed to be the same as language keywords, except for certain unprefixed function-names listed in A.4 Reserved Function Names.

In this specification the phrases must, must not, should, should not, may, required, and recommended, when used in normative text and rendered in small capitals, are to be interpreted as described in [RFC2119].

Certain aspects of language processing are described in this specification as implementation-defined or implementation-dependent.

  • [Definition: Implementation-defined indicates an aspect that may differ between implementations, but must be specified by the implementer for each particular implementation.]

  • [Definition: Implementation-dependent indicates an aspect that may differ between implementations, is not specified by this or any W3C specification, and is not required to be specified by the implementer for any particular implementation.]

A language aspect described in this specification as implementation-defined or implementation dependent may be further constrained by the specifications of a host language in which XPath is embedded.

2.1.1 Grammar Notation

Changes in 4.0  

  1. The EBNF notation has been extended to allow the constructs (A ++ ",") (one or more occurrences of A, comma-separated, and (A ** ",") (zero or more occurrences of A, comma-separated.   [Issue 1366 PR 1498 30 October 2024]

The grammar of XPath 4.0 is defined using a version of EBNF defined in A.1.1 Notation. The notation is based on the EBNF dialect used in the XML specification, with two notable additions derived from the Invisible XML grammar:

(A ++ ",") represents a sequence of one or more comma-separated occurrences of A.

(A ** ",") represents a sequence of zero or more comma-separated occurrences of A.

For example the following production rule indicates that an Expr consists of one or more occurrences of ExprSingle, separated by commas:

Expr ::= (ExprSingle ++ ",")

In principle any production can be used for the separator, but in practice this notation is only used in cases where the separator is a simple constant string.

EBNF grammar rules appear throughout the specification for ease of reference, and the entire grammar is summarized in A.1 EBNF. For XPath 4.0, the top-level production rule is XPath, representing an XPath expression.

2.1.2 Values

Changes in 4.0  

  1. The term atomic value has been replaced by atomic item.   [Issue 1337 PR 1361 2 August 2024]

[Definition: In the data model, a value is always a sequence.]

[Definition: A sequence is an ordered collection of zero or more items.]

[Definition: An item is either an atomic item, a node, or a function item.]

[Definition: An atomic item is a value in the value space of an atomic type, as defined in [XML Schema 1.0] or [XML Schema 1.1].]

[Definition: A node is an instance of one of the node kinds defined in Section 5 NodesDM40.] Each node has a unique node identity, a typed value, and a string value. In addition, some nodes have a name. The typed value of a node is a sequence of zero or more atomic items. The string value of a node is a value of type xs:string. The name of a node is a value of type xs:QName.

[Definition: A function item is an item that can be called using a dynamic function call.]

Maps (see 4.13.1 Maps) and arrays (see 4.13.2 Arrays) are specific kinds of function items.

[Definition: A sequence containing exactly one item is called a singleton.] An item is identical to a singleton sequence containing that item. Sequences are never nested—for example, combining the values 1, (2, 3), and ( ) into a single sequence results in the sequence (1, 2, 3). [Definition: A sequence containing zero items is called an empty sequence.]

[Definition: The term XDM instance is used, synonymously with the term value, to denote an unconstrained sequence of items.]

Element nodes have a property called in-scope namespaces. [Definition: The in-scope namespaces property of an element node is a set of namespace bindings, each of which associates a namespace prefix with a URI.] For a given element, one namespace binding may have an empty prefix; the URI of this namespace binding is the default namespace within the scope of the element.

In [XML Path Language (XPath) Version 1.0], the in-scope namespaces of an element node are represented by a collection of namespace nodes arranged on a namespace axis. As of XPath 2.0, the namespace axis is deprecated and need not be supported by a host language. A host language that does not support the namespace axis need not represent namespace bindings in the form of nodes.

2.1.3 Namespaces and QNames

[Definition: An expanded QName is a triple: its components are a prefix, a local name, and a namespace URI. In the case of a name in no namespace, the namespace URI and prefix are both absent. In the case of a name in the default namespace, the prefix is absent.] When comparing two expanded QNames, the prefixes are ignored: the local name parts must be equal under the Unicode codepoint collation (Section 5.3.1 CollationsFO40), and the namespace URI parts must either both be absent, or must be equal under the Unicode codepoint collation.

In the XPath 4.0 grammar, QNames representing the names of elements, attributes, functions, variables, types, or other such constructs are written as instances of the grammatical production EQName.

EQName ::= QName | URIQualifiedName
QName ::= [http://www.w3.org/TR/REC-xml-names/#NT-QName]Names
/* xgc: xml-version */
URIQualifiedName ::= BracedURILiteral NCName
/* ws: explicit */
BracedURILiteral ::= "Q" "{" [^{}]* "}"
/* ws: explicit */
NCName ::= [http://www.w3.org/TR/REC-xml-names/#NT-NCName]Names
/* xgc: xml-version */

The EQName production allows a QName to be written in one of three ways:

  • local-name only (for example, invoice).

    A name written in this form has no prefix, and the rules for determining the namespace depend on the context in which the name appears. This form is a lexical QName.

  • prefix plus local-name (for example, my:invoice).

    In this case the prefix and local name of the QName are as written, and the namespace URI is inferred from the prefix by examining the in-scope namespaces in the static context where the QName appears; the context must include a binding for the prefix. This form is a lexical QName.

  • URI plus local-name (for example, Q{http://example.com/ns}invoice).

    In this case the local name and namespace URI are as written, and the prefix is absent. This way of writing a QName is context-free, which makes it particularly suitable for use in expressions that are generated by software. This form is a URIQualifiedName. If the BracedURILiteral has no content (for example, Q{}invoice) then the namespace URI of the QName is absent.

[Definition: A lexical QName is a name that conforms to the syntax of the QName production].

The namespace URI value in a URIQualifiedName is whitespace normalized according to the rules for the xs:anyURI type in Section 3.2.17 anyURI XS1-2 or Section 3.3.17 anyURI XS11-2. It is a static error [err:XQST0070] if the namespace URI for an EQName is http://www.w3.org/2000/xmlns/.

Here are some examples of EQNames:

This document uses the following namespace prefixes to represent the namespace URIs with which they are listed. Although these prefixes are used within this specification to refer to the corresponding namespaces, not all of these bindings will necessarily be present in the static context of every expression, and authors are free to use different prefixes for these namespaces, or to bind these prefixes to different namespaces.

  • xs: http://www.w3.org/2001/XMLSchema

  • fn: http://www.w3.org/2005/xpath-functions

  • array: http://www.w3.org/2005/xpath-functions/array

  • map: http://www.w3.org/2005/xpath-functions/map

  • math: http://www.w3.org/2005/xpath-functions/math

  • err: http://www.w3.org/2005/xqt-errors (see 2.4.2 Identifying and Reporting Errors).

  • output: http://www.w3.org/2010/xslt-xquery-serialization

[Definition: Within this specification, the term URI refers to a Universal Resource Identifier as defined in [RFC3986] and extended in [RFC3987] with the new name IRI.] The term URI has been retained in preference to IRI to avoid introducing new names for concepts such as “Base URI” that are defined or referenced across the whole family of XML specifications.

Note:

In most contexts, processors are not required to raise errors if a URI is not lexically valid according to [RFC3986] and [RFC3987]. See 2.5.5 URI Literals for details.

2.2 Expression Context

[Definition: The expression context for a given expression consists of all the information that can affect the result of the expression.]

This information is organized into two categories called the static context and the dynamic context.

2.2.1 Static Context

Changes in 4.0  

  1. The default namespace for elements and types can be set to the value ##any, allowing unprefixed names in axis steps to match elements with a given local name in any namespace.   [Issue 296 PR 1181 30 April 2024]

  2. Parts of the static context that were there purely to assist in static typing, such as the statically known documents, were no longer referenced and have therefore been dropped.   [Issue 1343 PR 1344 23 September 2024]

  3. The context value static type, which was there purely to assist in static typing, has been dropped.   [Issue 1495 ]

[Definition: The static context of an expression is the information that is available during static analysis of the expression, prior to its evaluation.] This information can be used to decide whether the expression contains a static error.

The individual components of the static context are described below. A default initial value for each component must be specified by the host language. The scope of each component is specified in C.1 Static Context Components.

  • [Definition: XPath 1.0 compatibility mode. This value is true if rules for backward compatibility with XPath Version 1.0 are in effect; otherwise it is false. ]

  • [Definition: Statically known namespaces. This is a mapping from prefix to namespace URI that defines all the namespaces that are known during static processing of a given expression.]

    The URI value is whitespace normalized according to the rules for the xs:anyURI type in Section 3.2.17 anyURI XS1-2 or Section 3.3.17 anyURI XS11-2.

    The statically known namespaces may include a binding for the zero-length prefix; however, this is used only in limited circumstances because the rules for resolving unprefixed QNames depend on how such a name is used.

    Note the difference between in-scope namespaces, which is a dynamic property of an element node, and statically known namespaces, which is a static property of an expression.

  • [Definition: Default namespace for elements and types. This is either a namespace URI, or the special value "##any", or absentDM40. This indicates how unprefixed QNames are interpreted when they appear in a position where an element name or type name is expected.]

    • If the value is set to a namespace URI, this namespace is used for any such unprefixed QName. The URI value is whitespace-normalized according to the rules for the xs:anyURI type in Section 3.2.17 anyURI XS1-2 or Section 3.3.17 anyURI XS11-2.

    • The special value "##any" indicates that:

      • When an unprefixed QName is used as a name test for selecting named elements in an axis step, the name test will match an element having the specified local name, in any namespace or none.

      • When an unprefixed QName is used in a context where a type name is expected (but not as a function name), the default namespace is the xs namespace, http://www.w3.org/2001/XMLSchema.

      • In any other context, an unprefixed QName represents a name in no namespace.

    • If the value is absentDM40, an unprefixed QName representing an element or type name is interpreted as being in no namespace.

  • [Definition: Default function namespace. This is either a namespace URI, or absentDM40. The namespace URI, if present, is used for any unprefixed QName appearing in a position where a function name is expected.] The URI value is whitespace-normalized according to the rules for the xs:anyURI type in Section 3.2.17 anyURI XS1-2 or Section 3.3.17 anyURI XS11-2

    In its simplest form its value is simply a whitespace-normalized xs:anyURI value (most commonly, the URI http://www.w3.org/2005/xpath-functions) to be used as the default namespace for unprefixed function names. However, the use of a more complex algorithm is not precluded, for example an algorithm which searches multiple namespaces for a matching name.

  • [Definition: In-scope schema definitions is a generic term for all the element declarations, attribute declarations, and schema type definitions that are in scope during static analysis of an expression.] It includes the following three parts:

  • [Definition: In-scope variables. This is a mapping from expanded QName to type. It defines the set of variables that are available for reference within an expression. The expanded QName is the name of the variable, and the type is the static type of the variable.]

    An expression that binds a variable extends the in-scope variables, within the scope of the variable, with the variable and its type. Within the body of an inline function expression, the in-scope variables are extended by the names and types of the function parameters.

  • [Definition: In-scope named item types. This is a mapping from expanded QName to named item types.]

    [Definition: A named item type is an ItemType identified by an expanded QName.]

    Named item types serve two purposes:

    • They allow frequently used item types, especially complex item types such as record types, to be given simple names, to avoid repeating the definition every time it is used.

    • They allow the definition of recursive types, which are useful for describing recursive data structures such as lists and trees. For details see 3.2.8.4 Recursive Record Types.

    Note:

    Named item types can be defined in a host language such as XQuery 4.0 and in XSLT 4.0, but not in XPath 4.0 itself. They are available in XPath only if the host language provides the ability to define them.

  • [Definition: Statically known function definitions. This is a set of function definitions.]

    Function definitions are described in 2.2.1.1 Function Definitions.

  • [Definition: Statically known collations. This is an implementation-defined mapping from URI to collation. It defines the names of the collations that are available for use in processing expressions.] [Definition: A collation is a specification of the manner in which strings and URIs are compared and, by extension, ordered. For a more complete definition of collation, see Section 5.3 Comparison of stringsFO40.]

  • [Definition: Static Base URI. This is an absolute URI, used to resolve relative URIs during static analysis. ] For example, it is used to resolve module location URIs in XQuery, and the URIs in xsl:import and xsl:include in XSLT. If E is a subexpression of F then the Static Base URI of E is the same as the Static Base URI of F. There are no constructs in XPath that require resolution of relative URI references during static analysis.

    The Static Base URI is available during dynamic evaluation by use of the fn:static-base-uri function, and is used implicitly during dynamic evaluation by functions such as fn:doc. Relative URI references are resolved as described in 2.5.6 Resolving a Relative URI Reference.

    At execution time, relative URIs supplied to functions such as fn:doc are resolved against the Executable Base URI, which may or may not be the same as the Static Base URI.

  • [Definition: Statically known decimal formats. This is a mapping from QNames to decimal formats, with one default format that has no visible name, referred to as the unnamed decimal format. Each format is available for use when formatting numbers using the fn:format-number function.]

    Decimal formats are described in 2.2.1.2 Decimal Formats.

2.2.1.1 Function Definitions

[Definition: A function definition contains information used to evaluate a static function call, including the name, parameters, and return type of the function.]

The properties of a function definition include:

  • The function name, which is an expanded QName.

  • Parameter definitions, specifically:

    The names of the parameters must be distinct.

    [Definition: A function definition has an arity range, which is a range of consecutive non-negative integers. If the function definition has M required parameters and N optional parameters, then its arity range is from M to M+N inclusive.]

    The static context may contain several function definitions with the same name, but the arity ranges of two such function definitions must not overlap. For example, if two function definitions A and B have the same function name, then:

    • It is acceptable for A to have two required parameters and no optional parameters, while B has three required parameters and one optional parameter.

    • It is not acceptable for A to have one required parameter while B has three optional parameters.

    Note:

    Implementations must ensure that no two function definitions have the same expanded QName and overlapping arity ranges (even if the signatures are consistent).

    XQuery and XSLT enforce this rule by defining a static error if the rule is violated; but further constraints may be needed if an API allows external functions to be added to the static context.

  • A return type (a sequence type)

  • The function category, which is one of application, system, or external:

    • [Definition: Application functions are function definitions written in a host language such as XQuery or XSLT whose syntax and semantics are defined in this family of specifications. Their behavior (including the rules determining the static and dynamic context) follows the rules for such functions in the relevant host language specification.] The most common application functions are functions written by users in XQuery or XSLT.

    • [Definition: System functions include the functions defined in [XQuery and XPath Functions and Operators 4.0], functions defined by the specifications of a host language, constructor functions for atomic types, and any additional functions provided by the implementation. System functions are sometimes called built-in functions.]

      The behavior of system functions follows the rules given for the individual function in this family of specifications, or in the specification of the particular processor implementation. A system function may have behavior that depends on the static or dynamic context of the caller (for example, comparing strings using the default collation from the dynamic context of the caller). Such functions are said to be context dependent.

    • [Definition: External functions can be characterized as functions that are neither part of the processor implementation, nor written in a language whose semantics are under the control of this family of specifications. The semantics of external functions, including any context dependencies, are entirely implementation-defined. In XSLT, external functions are called Section 24.1 Extension Functions XT30. ]

      For example, an implementation might provide a mechanism allowing external functions to be written in a language such as Java or Python. The way in which argument and return values are converted between the XDM type system and the type system of the external language is implementation-defined.

    [Definition: A function definition is said to be context dependent if its result depends on the static or dynamic context of its caller. A function definition may be context-dependent for some arities in its arity range, and context-independent for others: for example fn:name#0 is context-dependent while fn:name#1 is context-independent.]

    Note:

    Some system functions, such as fn:position, fn:last, and fn:static-base-uri, exist for the sole purpose of providing information about the static or dynamic context of their caller.

    Note:

    Application functions are context dependent only to the extent that they define optional parameters with default values that are context dependent.

  • A (possibly empty) set of function annotations

  • A body. The function body contains the logic that enables the function result to be computed from the supplied arguments and information in the static and dynamic context.

The function definitions present in the static context are available for reference from a static function call, or from a named function reference.

2.2.1.2 Decimal Formats

Changes in 4.0  

  1. Several decimal format properties, including minus sign, exponent separator, percent, and per-mille, can now be rendered as arbitrary strings rather than being confined to a single character.   [Issue 1048 PR 1250 3 June 2024]

Each decimal format defines a set of properties, which control the interpretation of characters in the picture string supplied to the fn:format-number function, and also specify characters to be used in the result of formatting the number.

The following properties specify characters used both in the picture string, and in the formatted number. In each case the value is a single character:

  • [Definition: decimal-separator is the character used to separate the integer part of the number from the fractional part, both in the picture string and in the formatted number; the default value is U+002E (FULL STOP, PERIOD, .) .]

  • [Definition: exponent-separator is the character used to separate the mantissa from the exponent in scientific notation both in the picture string and in the formatted number; the default value is U+0065 (LATIN SMALL LETTER E, e) .]

  • [Definition: grouping-separator is the character typically used as a thousands separator, both in the picture string and in the formatted number; the default value is U+002C (COMMA, ,) .]

  • [Definition: percent is the character used both in the picture string and in the formatted number to indicate that the number is written as a per-hundred fraction; the default value is U+0025 (PERCENT SIGN, %) .]

  • [Definition: per-mille is the character used both in the picture string and in the formatted number to indicate that the number is written as a per-thousand fraction; the default value is U+2030 (PER MILLE SIGN, ) .]

  • [Definition: zero-digit is the character used to represent the digit zero; the default value is U+0030 (DIGIT ZERO, 0) . This character must be a digit (category Nd in the Unicode property database), and it must have the numeric value zero. This property implicitly defines the ten Unicode characters that are used to represent the values 0 to 9: Unicode is organized so that each set of decimal digits forms a contiguous block of characters in numerical sequence. Within the picture string any of these ten character can be used (interchangeably) as a place-holder for a mandatory digit. Within the final result string, these ten characters are used to represent the digits zero to nine.]

In the case of the the properties decimal-separator, grouping-separator, exponent-separator, percent and per-mille, the property may take the form m:r, where m is a single-character marker used in the picture string to indicate where the relevant output should appear, and r is the string used to represent the property in the result. For example, setting the percent property to "%:pc" means that the value 0.10, formatted with the picture string #0%, results in the output 10pc.

The following properties specify characters to be used in the picture string supplied to the fn:format-number function, but not in the formatted number. In each case the value must be a single character.

  • [Definition: digit is a character used in the picture string to represent an optional digit; the default value is U+0023 (NUMBER SIGN, #) .]

  • [Definition: pattern-separator is a character used to separate positive and negative sub-pictures in a picture string; the default value is U+003B (SEMICOLON, ;) .]

The following properties specify characters or strings that may appear in the result of formatting the number, but not in the picture string:

  • [Definition: infinity is the string used to represent the double value infinity (INF); the default value is the string "Infinity" ]

  • [Definition: NaN is the string used to represent the double value NaN (not a number); the default value is the string "NaN" ]

  • [Definition: minus-sign is the single character used to mark negative numbers; the default value is U+002D (HYPHEN-MINUS, -) .]

2.2.2 Dynamic Context

Changes in 4.0  

  1. The concept of the context item has been generalized, so it is now a context value. That is, it is no longer constrained to be a single item.   [Issue 129 PR 368 14 September 2023]

  2. The rules regarding the document-uri property of nodes returned by the fn:collection function have been relaxed.   [Issue 1161 PR 1265 11 June 2024]

[Definition: The dynamic context of an expression is defined as information that is needed for the dynamic evaluation of an expression.] If evaluation of an expression relies on some part of the dynamic context that is absentDM40, a type error is raised [err:XPDY0002].

Note:

In previous versions of the specification, this was classified as a dynamic error. The change allows the error to be raised during static analysis when possible; for example a function written as fn($x) { @code } can now be reported as an error whether or not the function is actually evaluated. The actual error code remains unchanged for backwards compatibility reasons.

There are other cases where static detection of the error is not possible.

The individual components of the dynamic context are described below. Further rules governing the semantics of these components can be found in C.2 Dynamic Context Components.

The dynamic context consists of all the components of the static context, and the additional components listed below.

[Definition: The first three components of the dynamic context (context value, context position, and context size) are called the focus of the expression. ] The focus enables the processor to keep track of which items are being processed by the expression. If any component in the focus is defined, both the context value and context position are known.

Note:

If any component in the focus is defined, context size is usually defined as well. However, when streaming, the context size cannot be determined without lookahead, so it may be undefined. If so, expressions like last() will raise a dynamic error because the context size is undefined.

[Definition: A fixed focus is a focus for an expression that is evaluated once, rather than being applied to a series of values; in a fixed focus, the context value is set to one specific value, the context position is 1, and the context size is 1.]

[Definition: A singleton focus is a fixed focus in which the context value is a singleton item.]. With a singleton focus, the context value is a single item, the context position is 1, and the context size is 1.

Certain language constructs, notably the path operator E1/E2, the simple map operator E1!E2, and the predicate E1[E2], create a new focus for the evaluation of a sub-expression. In these constructs, E2 is evaluated once for each item in the sequence that results from evaluating E1. Each time E2 is evaluated, it is evaluated with a different focus. The focus for evaluating E2 is referred to below as the inner focus, while the focus for evaluating E1 is referred to as the outer focus. The inner focus is used only for the evaluation of E2. Evaluation of E1 continues with its original focus unchanged.

  • [Definition: The context value is the value currently being processed.] In many cases (but not always), the context value will be a single item. [Definition: When the context value is a single item, it can also be referred to as the context item; when it is a single node, it can also be referred to as the context node.] The context value is returned by an expression consisting of a single dot (.). When an expression E1/E2 or E1[E2] is evaluated, each item in the sequence obtained by evaluating E1 becomes the context value in the inner focus for an evaluation of E2.

  • [Definition: The context position is the position of the context value within the series of values currently being processed.] It changes whenever the context value changes. When the focus is defined, the value of the context position is an integer greater than zero. The context position is returned by the expression fn:position(). When an expression E1/E2 or E1[E2] is evaluated, the context position in the inner focus for an evaluation of E2 is the position of the context value in the sequence obtained by evaluating E1. The position of the first item in a sequence is always 1 (one). The context position is always less than or equal to the context size.

  • [Definition: The context size is the number of values in the series of values currently being processed.] Its value is always an integer greater than zero. The context size is returned by the expression fn:last(). When an expression E1/E2 or E1[E2] is evaluated, the context size in the inner focus for an evaluation of E2 is the number of items in the sequence obtained by evaluating E1.

  • [Definition: Variable values. This is a mapping from expanded QName to value. It contains the same expanded QNames as the in-scope variables in the static context for the expression. The expanded QName is the name of the variable and the value is the dynamic value of the variable, which includes its dynamic type.]

  • [Definition: Dynamically known function definitions. This is a set of function definitions. It includes the statically known function definitions as a subset, but may include other function definitions that are not known statically. ]

    The function definitions in the dynamic context are used primarily by the fn:function-lookup function.

    If two function definitions in the dynamically known function definitions have the same name, then their arity ranges must not overlap.

    Note:

    The reason for allowing named functions to be available dynamically beyond those that are available statically is primarily to allow for cases where the run-time execution environment is significantly different from the compile-time environment. This could happen, for example, if a stylesheet or query is compiled within a web server and then executed in the web browser. The fn:function-lookup function allows dynamic discovery of resources that were not available statically.

  • [Definition: Current dateTime. This information represents an implementation-dependent point in time during the processing of an expression, and includes an explicit timezone. It can be retrieved by the fn:current-dateTime function. If called multiple times during the execution of an expression, this function always returns the same result.]

  • [Definition: Implicit timezone. This is the timezone to be used when a date, time, or dateTime value that does not have a timezone is used in a comparison or arithmetic operation. The implicit timezone is an implementation-defined value of type xs:dayTimeDuration. See Section 3.2.7.3 Timezones XS1-2 or Section 3.3.7 dateTime XS11-2 for the range of valid values of a timezone.]

  • [Definition: Executable Base URI. This is an absolute URI used to resolve relative URIs during the evaluation of expressions; it is used, for example, to resolve a relative URI supplied to the fn:doc or fn:unparsed-text functions. ]

    URIs are resolved as described in 2.5.6 Resolving a Relative URI Reference.

    The function fn:static-base-uri, despite its name, returns the value of the Executable Base URI.

    In many straightforward processing scenarios, the Executable Base URI in the dynamic context will be the same as the Static Base URI for the corresponding expression in the static context. There are situations, however, where they may differ:

    • Some processors may allow the static analysis of a query or stylesheet to take place on a development machine, while execution of the query or stylesheet happens on a test or production server. In this situation, resources needed during static analysis (such as other modules of the query or stylesheet) will be located on the development machine, by reference to the Static Base URI, while resources needed during execution (such as reference data files) will be located on the production machine, accessed via the Executable Base URI.

    • When the fn:static-base-uri function is called within the initializing expression of an optional parameter in a function declaration, it returns the executable base URI of the relevant function call. This allows a user-written function to accept two parameters: a required parameter containing a relative URI, and an optional parameter containing a base URI. The optional parameter can be given a default value of fn:static-base-uri(), allowing the code in the function body to resolve the relative URI against the executable base URI of the caller.

  • [Definition: Default collation. This identifies one of the collations in statically known collations as the collation to be used by functions and operators for comparing and ordering values of type xs:string and xs:anyURI (and types derived from them) when no explicit collation is specified.]

    Note:

    Although the default collation is defined (in 4.0) as a property of the dynamic context, its value will in nearly all cases be known statically. The reason it is defined in the dynamic context is to allow a call on the fn:default-collation function to be used when defining the default value of an optional parameter to a user-defined function. In this situation, the actual value supplied for the parameter is taken from the dynamic context of the relevant function call.

  • [Definition: Default language. This is the natural language used when creating human-readable output (for example, by the functions fn:format-date and fn:format-integer) if no other language is requested. The value is a language code as defined by the type xs:language.]

  • [Definition: Default calendar. This is the calendar used when formatting dates in human-readable output (for example, by the functions fn:format-date and fn:format-dateTime) if no other calendar is requested. The value is a string.]

  • [Definition: Default place. This is a geographical location used to identify the place where events happened (or will happen) when processing dates and times using functions such as fn:format-date, fn:format-dateTime, and fn:civil-timezone, if no other place is specified. It is used when translating timezone offsets to civil timezone names, and when using calendars where the translation from ISO dates/times to a local representation is dependent on geographical location. Possible representations of this information are an ISO country code or an Olson timezone name, but implementations are free to use other representations from which the above information can be derived. The only requirement is that it should uniquely identify a civil timezone, which means that country codes for countries with multiple timezones, such as the United States, are inadequate.]

  • [Definition: Available documents. This is a mapping of strings to document nodes. Each string represents the absolute URI of a resource. The document node is the root of a tree that represents that resource using the data model. The document node is returned by the fn:doc function when applied to that URI.] The set of available documents may be empty.

  • [Definition: Available text resources. This is a mapping of strings to text resources. Each string represents the absolute URI of a resource. The resource is returned by the fn:unparsed-text function when applied to that URI.] The set of available text resources may be empty.

  • [Definition: Available collections. This is a mapping of strings to sequences of items. Each string represents the absolute URI of a resource. The sequence of items represents the result of the fn:collection function when that URI is supplied as the argument. ] The set of available collections may be empty.

    Ideally, for every document node D that is in the target of a mapping in available item collections, or that is the root of a tree containing such a node, the document-uri property of D should either be absent, or should be a URI U such that available documents contains a mapping from U to D.

    Note:

    That is to say, the document-uri property of nodes returned by the fn:collection function should be such that calling fn:doc with that URI returns the relevant node.

    It is not always possible to ensure this, especially in cases where dereferencing of document or collection URIs is configurable using configuration files or user-supplied resolver code.

  • [Definition: Default collection. This is the sequence of items that would result from calling the fn:collection function with no arguments.] The value of default collection may be initialized by the implementation.

  • [Definition: Available URI collections. This is a mapping of strings to sequences of URIs. The string represents the absolute URI of a resource which can be interpreted as an aggregation of a number of individual resources each of which has its own URI. The sequence of URIs represents the result of the fn:uri-collection function when that URI is supplied as the argument. ] There is no implication that the URIs in this sequence can be successfully dereferenced, or that the resources they refer to have any particular media type.

    Note:

    An implementation may maintain some consistent relationship between the available collections and the available URI collections, for example by ensuring that the result of fn:uri-collection(X)!fn:doc(.) is the same as the result of fn:collection(X). However, this is not required. The fn:uri-collection function is more general than fn:collection in that it allows access to resources other than XML documents; at the same time, fn:collection allows access to nodes that might lack individual URIs, for example nodes corresponding to XML fragments stored in the rows of a relational database.

  • [Definition: Default URI collection. This is the sequence of URIs that would result from calling the fn:uri-collection function with no arguments.] The value of default URI collection may be initialized by the implementation.

  • [Definition: Environment variables. This is a mapping from names to values. Both the names and the values are strings. The names are compared using an implementation-defined collation, and are unique under this collation. The set of environment variables is implementation-defined and may be empty.]

    Note:

    A possible implementation is to provide the set of POSIX environment variables (or their equivalent on other operating systems) appropriate to the process in which the expression is evaluated.

2.3 Processing Model

Changes in 4.0  

  1. The static typing option has been dropped.   [Issue 1343 PR 1344 3 September 2024]

XPath 4.0 is defined in terms of the data model and the expression context.

Processing                          Model Overview

Figure 1: Processing Model Overview

Figure 1 provides a schematic overview of the processing steps that are discussed in detail below. Some of these steps are completely outside the domain of XPath 4.0; in Figure 1, these are depicted outside the line that represents the boundaries of the language, an area labeled external processing. The external processing domain includes generation of XDM instances that represent the data to be queried (see 2.3.1 Data Model Generation), schema import processing (see 2.3.2 Schema Import Processing), and serialization. The area inside the boundaries of the language is known as the XPath processing domain , which includes the static analysis and dynamic evaluation phases (see 2.3.3 Expression Processing). Consistency constraints on the XPath processing domain are defined in 2.3.5 Consistency Constraints.

2.3.1 Data Model Generation

The input data for an expression must be represented as one or more XDM instances. This process occurs outside the domain of XPath 4.0, which is why Figure 1 represents it in the external processing domain.

In many cases the input data might originate as XML. Here are some steps by which an XML document might be converted to an XDM instance:

  1. A document may be parsed using an XML parser that generates an XML Information Set (see [XML Infoset]). The parsed document may then be validated against one or more schemas. This process, which is described in [XML Schema 1.0 Part 1] or [XML Schema 1.1 Part 1], results in an abstract information structure called the Post-Schema Validation Infoset (PSVI). If a document has no associated schema, its Information Set is preserved. (See DM1 in Figure 1)

  2. The Information Set or PSVI may be transformed into an XDM instance by a process described in [XQuery and XPath Data Model (XDM) 4.0]. (See DM2 in Figure 1)

The above steps provide an example of how an XDM instance might be constructed. An XDM instance might also be constructed in some other way (see DM3 in Figure 1), for example it might be synthesized directly from a relational database, or derived by parsing a JSON text or a CSV file. Whatever the origin, XPath 4.0 is defined in terms of the data model, but it does not place any constraints on how XDM instances are constructed.

The remainder of this section is concerned with the common case where XML data is being processed.

[Definition: Each element node and attribute node in an XDM instance has a type annotation (described in Section 2.8 Schema InformationDM40). The type annotation of a node is a reference to a schema type. ] The type-name of a node is the name of the type referenced by its type annotation (but note that the type annotation can be a reference to an anonymous type). If the XDM instance was derived from a validated XML document as described in Section 3.3 Construction from a PSVIDM40, the type annotations of the element and attribute nodes are derived from schema validation. XPath 4.0 does not provide a way to directly access the type annotation of an element or attribute node.

The value of an attribute is represented directly within the attribute node. An attribute node whose type is unknown (such as might occur in a schemaless document) is given the type annotation xs:untypedAtomic.

The value of an element is represented by the children of the element node, which may include text nodes and other element nodes. The type annotation of an element node indicates how the values in its child text nodes are to be interpreted. An element that has not been validated (such as might occur in a schemaless document) is annotated with the schema type xs:untyped. An element that has been validated and found to be partially valid is annotated with the schema type xs:anyType. If an element node is annotated as xs:untyped, all its descendant element nodes are also annotated as xs:untyped. However, if an element node is annotated as xs:anyType, some of its descendant element nodes may have a more specific type annotation.

2.3.2 Schema Import Processing

The in-scope schema definitions in the static context are provided by the host language (see step SI1 in Figure 1) and must satisfy the consistency constraints defined in 2.3.5 Consistency Constraints.

2.3.3 Expression Processing

XPath 4.0 defines two phases of processing called the static analysis phase and the dynamic evaluation phase (see Figure 1). During the static analysis phase, static errors, dynamic errors, or type errors may be raised. During the dynamic evaluation phase, only dynamic errors or type errors may be raised. These kinds of errors are defined in 2.4.1 Kinds of Errors.

Within each phase, an implementation is free to use any strategy or algorithm whose result conforms to the specifications in this document.

2.3.3.1 Static Analysis Phase

[Definition: The static analysis phase depends on the expression itself and on the static context. The static analysis phase does not depend on input data (other than schemas).]

During the static analysis phase, the XPath expression is typically parsed into an internal representation called the operation tree (step SQ1 in Figure 1). A parse error is raised as a static error [err:XPST0003]. The static context is initialized by the implementation (step SQ2). The static context is used to resolve schema type names, function names, namespace prefixes, and variable names (step SQ4). If a name of one of these kinds in the operation tree is not found in the static context, a static error ([err:XPST0008] or [err:XPST0017]) is raised (however, see exceptions to this rule in 3.2.7.2 Element Test and 3.2.7.4 Attribute Test.)

The operation tree is then typically normalized by making explicit the implicit operations such as atomization and extraction of effective boolean values (step SQ5).

During the static analysis phase, a processor may perform type analysis. The effect of type analysis is to assign a static type to each expression in the operation tree. [Definition: The static type of an expression is the best inference that the processor is able to make statically about the type of the result of the expression.] This specification does not define the rules for type analysis nor the static types that are assigned to particular expressions: the only constraint is that the inferred type must match all possible values that the expression is capable of returning.

Examples of inferred static types might be:

  • For the expression concat(a,b) the inferred static type is xs:string

  • For the expression $a = $v the inferred static type is xs:boolean

  • For the expression $s[exp] the inferred static type has the same item type as the static type of $s, but a cardinality that allows the empty sequence even if the static type of $s does not allow an empty sequence.

  • The inferred static type of the expression data($x) (whether written explicitly or inserted into the operation tree in places where atomization is implicit) depends on the inferred static type of $x: for example, if $x has type element(*, xs:integer) then data($x) has static type xs:integer.

In XQuery 1.0 and XPath 2.0, rules for static type inferencing were published normatively in [XQuery 1.0 and XPath 2.0 Formal Semantics], but implementations were allowed to refine these rules to infer a more precise type where possible. In subsequent versions, the rules for static type inferencing are entirely implementation-dependent.

Every kind of expression also imposes requirements on the type of its operands. For example, with the expression substring($a, $b, $c), $a must be of type xs:string (or something that can be converted to xs:string by the function calling rules), while $b and $c must be numeric.

A processor may raise a type error during static analysis if the inferred static type of an expression has no overlap (intersection) with the required type, and cannot be converted to the required type using the coercion rules. For example, given the call fn:upper-case($s), the processor may raise an error if the declared or inferred type of $s is xs:integer, but not if it is xs:anyAtomicType.

In addition, type analysis may conclude that an expression is implausible. Implausible expressions may be considered erroneous unless such checks have been disabled. This topic is described further in 2.4.6 Implausible Expressions.

Alternatively, the processor may defer all type checking until the dynamic evaluation phase.

2.3.3.2 Dynamic Evaluation Phase

[Definition: The dynamic evaluation phase is the phase during which the value of an expression is computed.] It is dependent on successful completion of the static analysis phase.

The dynamic evaluation phase can occur only if no errors were detected during the static analysis phase.

The dynamic evaluation phase depends on the operation tree of the expression being evaluated (step DQ1), on the input data (step DQ4), and on the dynamic context (step DQ5), which in turn draws information from the external environment (step DQ3) and the static context (step DQ2). The dynamic evaluation phase may create new data-model values (step DQ4) and it may extend the dynamic context (step DQ5)—for example, by binding values to variables.

[Definition: Every value matches one or more sequence types. A value is said to have a dynamic type T if it matches (or is an instance of) the sequence type T.]

In many cases (but not all), one of the dynamic types that a value matches will be a subtype of all the others, in which case it makes sense to speak of “the dynamic type” of the value as meaning this single most specific type. In other cases (examples are empty maps and empty arrays) none of the dynamic types is more specific than all the others.

Note:

An atomic item has a type annotation which will always be a subtype of all the other types that it matches; we can therefore refer to this as the dynamic type of the atomic item without ambiguity.

A value may match a dynamic type that is more specific than the static type of the expression that computed it (for example, the static type of an expression might be xs:integer*, denoting a sequence of zero or more integers, but at evaluation time its value may be an instance of xs:integer, denoting exactly one integer).

If an operand of an expression does not have a dynamic type that is a subtype of the static type required for that operand, a type error is raised [err:XPTY0004].

Even though static typing can catch many type errors before an expression is executed, it is possible for an expression to raise an error during evaluation that was not detected by static analysis. For example, an expression may contain a cast of a string into an integer, which is statically valid. However, if the actual value of the string at run time cannot be cast into an integer, a dynamic error will result. Similarly, an expression may apply an arithmetic operator to a value whose static type is xs:untypedAtomic. This is not a static error, but at run time, if the value cannot be successfully cast to a numeric type, a dynamic error will be raised.

2.3.4 Input Sources

XPath 4.0 has a set of functions that provide access to XML documents (fn:doc, fn:doc-available), collections (fn:collection, fn:uri-collection), text files (fn:unparsed-text, fn:unparsed-text-lines, fn:unparsed-text-available), and environment variables (fn:environment-variable, fn:available-environment-variables). These functions are defined in Section 13.6 Functions giving access to external informationFO40.

An expression can access input data either by calling one of these input functions or by referencing some part of the dynamic context that is initialized by the external environment, such as a variable or context value.

Note:

The EXPath Community Group has developed a File Module, which some implementations use to perform file system related operations such as reading or writing files and directories. Multiple files can be read or written from a single query.

2.3.5 Consistency Constraints

In order for XPath 4.0 to be well defined, the input XDM instances, the static context, and the dynamic context must be mutually consistent. The consistency constraints listed below are prerequisites for correct functioning of an XPath 4.0 implementation. Enforcement of these consistency constraints is beyond the scope of this specification. This specification does not define the result of an expression under any condition in which one or more of these constraints is not satisfied.

2.4 Error Handling

2.4.1 Kinds of Errors

As described in 2.3.3 Expression Processing, XPath 4.0 defines a static analysis phase, which does not depend on input data, and a dynamic evaluation phase, which does depend on input data. Errors may be raised during each phase.

[Definition: An error that can be detected during the static analysis phase, and is not a type error, is a static error.] A syntax error is an example of a static error.

[Definition: A dynamic error is an error that must be detected during the dynamic evaluation phase and may be detected during the static analysis phase.] Numeric overflow is an example of a dynamic error.

[Definition: A type error may be raised during the static analysis phase or the dynamic evaluation phase. During the static analysis phase, a type error occurs when the static type of an expression does not match the expected type of the context in which the expression occurs. During the dynamic evaluation phase, a type error occurs when the dynamic type of a value does not match the expected type of the context in which the value occurs.]

The outcome of the static analysis phase is either success or one or more type errors, static errors, or statically detected dynamic errors. The result of the dynamic evaluation phase is either a result value, a type error, or a dynamic error.

If more than one error is present, or if an error condition comes within the scope of more than one error defined in this specification, then any non-empty subset of these errors may be reported.

If an implementation can determine during the static analysis phase that an XPath expression, if evaluated, would necessarily raise a dynamic error or that an expression, if evaluated, would necessarily raise a type error, the implementation may (but is not required to) report that error during the static analysis phase.

An implementation can raise a dynamic error for an XPath expression statically only if the expression can never execute without raising that error, as in the following example:

error()

The following example contains a type error, which can be reported statically even if the implementation can not prove that the expression will actually be evaluated.

if (empty($arg))
then "cat" * 2
else 0

[Definition: In addition to static errors, dynamic errors, and type errors, an XPath 4.0 implementation may raise warnings, either during the static analysis phase or the dynamic evaluation phase. The circumstances in which warnings are raised, and the ways in which warnings are handled, are implementation-defined.]

In addition to the errors defined in this specification, an implementation may raise a dynamic error for a reason beyond the scope of this specification. For example, limitations may exist on the maximum numbers or sizes of various objects. An error must be raised if such a limitation is exceeded [err:XPDY0130].

2.4.2 Identifying and Reporting Errors

The errors defined in this specification are identified by QNames that have the form err:XPYYnnnn , where:

  • err denotes the namespace for XPath and XQuery errors, http://www.w3.org/2005/xqt-errors. This binding of the namespace prefix err is used for convenience in this document, and is not normative.

  • XP identifies the error as an XPath error (some errors, originally defined by XQuery and later added to XPath, use the code XQ instead).

  • YY denotes the error category, using the following encoding:

    • ST denotes a static error.

    • DY denotes a dynamic error.

    • TY denotes a type error.

  • nnnn is a unique numeric code.

Note:

The namespace URI for XPath and XQuery errors is not expected to change from one version of XPath to another. However, the contents of this namespace may be extended to include additional error definitions.

The method by which an XPath 4.0 processor reports error information to the external environment is implementation-defined.

An error can be represented by a URI reference that is derived from the error QName as follows: an error with namespace URI NS and local part LP can be represented as the URI reference NS # LP . For example, an error whose QName is err:XPST0017 could be represented as http://www.w3.org/2005/xqt-errors#XPST0017.

Note:

Along with a code identifying an error, implementations may wish to return additional information, such as the location of the error or the processing phase in which it was detected. If an implementation chooses to do so, then the mechanism that it uses to return this information is implementation-defined.

2.4.3 Handling Dynamic Errors

Except as noted in this document, if any operand of an expression raises a dynamic error, the expression also raises a dynamic error. If an expression can validly return a value or raise a dynamic error, the implementation may choose to return the value or raise the dynamic error (see 2.4.4 Errors and Optimization). For example, the logical expression expr1 and expr2 may return the value false if either operand returns false, or may raise a dynamic error if either operand raises a dynamic error.

If more than one operand of an expression raises an error, the implementation may choose which error is raised by the expression. For example, in this expression:

($x div $y) + xs:decimal($z)

both the sub-expressions ($x div $y) and xs:decimal($z) may raise an error. The implementation may choose which error is raised by the + expression. Once one operand raises an error, the implementation is not required, but is permitted, to evaluate any other operands.

[Definition: In addition to its identifying QName, a dynamic error may also carry a descriptive string and one or more additional values called error values.] An implementation may provide a mechanism whereby an application-defined error handler can process error values and produce diagnostic messages. The host language may also provide error handling mechanisms.

A dynamic error may be raised by a system function or operator. For example, the div operator raises an error if its operands are xs:decimal values and its second operand is equal to zero. Errors raised by system functions and operators are defined in [XQuery and XPath Functions and Operators 4.0] or the host language.

A dynamic error can also be raised explicitly by calling the fn:error function, which always raises a dynamic error and never returns a value. This function is defined in Section 3.1.1 fn:errorFO40. For example, the following function call raises a dynamic error, providing a QName that identifies the error, a descriptive string, and a diagnostic value (assuming that the prefix app is bound to a namespace containing application-defined error codes):

error(xs:QName("app:err057"), "Unexpected value", string($v))

2.4.4 Errors and Optimization

Because different implementations may choose to evaluate or optimize an expression in different ways, certain aspects of raising dynamic errors are implementation-dependent, as described in this section.

An implementation is always free to evaluate the operands of an operator in any order.

In some cases, a processor can determine the result of an expression without accessing all the data that would be implied by the formal expression semantics. For example, the formal description of filter expressions suggests that $s[1] should be evaluated by examining all the items in sequence $s, and selecting all those that satisfy the predicate position()=1. In practice, many implementations will recognize that they can evaluate this expression by taking the first item in the sequence and then exiting. If $s is defined by an expression such as //book[author eq 'Berners-Lee'], then this strategy may avoid a complete scan of a large document and may therefore greatly improve performance. However, a consequence of this strategy is that a dynamic error or type error that would be detected if the expression semantics were followed literally might not be detected at all if the evaluation exits early. In this example, such an error might occur if there is a book element in the input data with more than one author subelement.

The extent to which a processor may optimize its access to data, at the cost of not raising errors, is defined by the following rules.

Consider an expression Q that has an operand (sub-expression) E. In general the value of E is a sequence. At an intermediate stage during evaluation of the sequence, some of its items will be known and others will be unknown. If, at such an intermediate stage of evaluation, a processor is able to establish that there are only two possible outcomes of evaluating Q, namely the value V or an error, then the processor may deliver the result V without evaluating further items in the operand E. For this purpose, two values are considered to represent the same outcome if their items are pairwise the same, where nodes are the same if they have the same identity, and values are the same if they are equal and have exactly the same type.

There is an exception to this rule: If a processor evaluates an operand E (wholly or in part), then it is required to establish that the actual value of the operand E does not violate any constraints on its cardinality. For example, the expression $e eq 0 results in a type error if the value of $e contains two or more items. A processor is not allowed to decide, after evaluating the first item in the value of $e and finding it equal to zero, that the only possible outcomes are the value true or a type error caused by the cardinality violation. It must establish that the value of $e contains no more than one item.

These rules apply to all the operands of an expression considered in combination: thus if an expression has two operands E1 and E2, it may be evaluated using any samples of the respective sequences that satisfy the above rules.

The rules cascade: if A is an operand of B and B is an operand of C, then the processor needs to evaluate only a sufficient sample of B to determine the value of C, and needs to evaluate only a sufficient sample of A to determine this sample of B.

The effect of these rules is that the processor is free to stop examining further items in a sequence as soon as it can establish that further items would not affect the result except possibly by causing an error. For example, the processor may return true as the result of the expression S1 = S2 as soon as it finds a pair of equal values from the two sequences.

Another consequence of these rules is that where none of the items in a sequence contributes to the result of an expression, the processor is not obliged to evaluate any part of the sequence. Again, however, the processor cannot dispense with a required cardinality check: if an empty sequence is not permitted in the relevant context, then the processor must ensure that the operand is not an empty sequence.

Examples:

  • If an implementation can find (for example, by using an index) that at least one item returned by $expr1 in the following example has the value 47, it is allowed to return true as the result of the some expression, without searching for another item returned by $expr1 that would raise an error if it were evaluated.

    some $x in $expr1 satisfies $x = 47
  • In the following example, if an implementation can find (for example, by using an index) the product element-nodes that have an id child with the value 47, it is allowed to return these nodes as the result of the path expression, without searching for another product node that would raise an error because it has an id child whose value is not an integer.

    //product[id = 47]

For a variety of reasons, including optimization, implementations may rewrite expressions into a different form. There are a number of rules that limit the extent of this freedom:

  • Other than the raising or not raising of errors, the result of evaluating a rewritten expression must conform to the semantics defined in this specification for the original expression.

    Note:

    This allows an implementation to return a result in cases where the original expression would have raised an error, or to raise an error in cases where the original expression would have returned a result. The main cases where this is likely to arise in practice are (a) where a rewrite changes the order of evaluation, such that a subexpression causing an error is evaluated when the expression is written one way and is not evaluated when the expression is written a different way, and (b) where intermediate results of the evaluation cause overflow or other out-of-range conditions.

    Note:

    This rule does not mean that the result of the expression will always be the same in non-error cases as if it had not been rewritten, because there are many cases where the result of an expression is to some degree implementation-dependent or implementation-defined.

  • The rules described in 2.4.5 Guarded Expressions ensure that for certain kinds of expression (for example conditional expressions), changing the order of evaluation of subexpressions does not result in dynamic errors that would not otherwise occur.

  • Expressions must not be rewritten in such a way as to create or remove static errors. The static errors in this specification are defined for the original expression, and must be preserved if the expression is rewritten.

  • As stated earlier, an expression must not be rewritten to dispense with a required cardinality check: for example, string-length(//title) must raise an error if the document contains more than one title element.

2.4.5 Guarded Expressions

Changes in 4.0  

  1. The rules for “errors and optimization” have been tightened up to disallow many cases of optimizations that alter error behavior. In particular there are restrictions on reordering the operands of and and or, and of predicates in filter expressions, in a way that might allow the processor to raise dynamic errors that the author intended to prevent.   [Issue 71 PR 230 15 November 2022]

[Definition: An expression E is said to be guarded by some governing condition C if evaluation of E is not allowed to fail with a dynamic error except when C applies.]

For example, in a conditional expression if (P) then T else F, the subexpression T is guarded by P, and the subexpression F is guarded by not(P). One way an implementation can satisfy this rule is by not evaluating T unless P is true, and likewise not evaluating F unless P is false. Another way of satisfying the rule is for the implementation to evaluate all the subexpressions, but to catch any errors that occur in a guarded subexpression so they are not propagated.

The existence of this rule enables errors to be prevented by writing expressions such as if ($y eq 0) then "N/A" else ($x div $y). This example will never fail with a divide-by-zero error because the else branch of the conditional is guarded.

Similarly, in the mapping expression E1!E2, the subexpression E2 is guarded by the existence of an item from E1. This means, for example, that the expression (1 to $n)!doc('bad.xml') must not raise a dynamic error if $n is zero. The rule governing evaluation of guarded expressions is phrased so as not to disallow “loop-lifting” or “constant-folding” optimizations whose aim is to avoid repeated evaluation of a common subexpression; but such optimizations must not result in errors that would not otherwise occur.

The complete list of expressions that have guarded subexpressions is as follows:

  • In a conditional expression (IfExpr) the then branch is guarded by the condition being true, and the else branch is guarded by the condition being false.

  • In an and expression (AndExpr), the second operand is guarded by the value of the first operand being true.

  • In an or expression (OrExpr), the second operand is guarded by the value of the first operand being false.

  • In an otherwise expression (OtherwiseExpr), the second operand is guarded by the value of the first operand being an empty sequence.

  • In a path expression of the form E1/E2 or E1//E2, and in a mapping expression of the form E1!E2, the right-hand operand E2 is guarded by the existence of at least one item in the result of evaluating E1.

    This rule applies even if E2 does not reference the context value. For example, no dynamic error can be thrown by the expression (1 to $n)!doc('bad.xml') in the case where $n is zero.

  • In a filter expression of the form E[P], the predicate P is guarded by the existence of at least one item in the result of evaluating E.

    This rule has the consequence that in a filter expression with multiple predicates, such as E[P1][P2], evaluation of P2 must not raise a dynamic error unless P1 returns true. This rule does not prevent reordering of predicates (for example, to take advantage of indexes), but it does require that any such reordering must not result in errors that would not otherwise occur.

  • In a for expression (ForExpr) such as for $x in S return E, the expression E is guarded by the existence of an item bound to $x.

    This means that the expression for $x in 1 to $n return doc('bad.xml') must not raise a dynamic error in the case where $n is zero.

  • In a quantified expression (QuantifiedExpr) such as some $x in S satisfies P, the expression P is guarded by the existence of an item bound to $x.

The fact that an expression is guarded does not remove the obligation to report static errors in the expression; nor does it remove the option to report statically detectable type errors.

Note:

These rules do not constrain the order of evaluation of subexpressions. For example, given an expression such as //person[@first = "Winston"][@last = "Churchill"], or equivalently //person[@first = "Winston" and @last = "Churchill"], an implementation might use an index on the value of @last to select items that satisfy the second condition, and then filter these items on the value of the first condition. Alternatively, it might evaluate both predicates in parallel. Or it might interpose an additional redundant condition: //person[string-length(@first) + string-length(@last) = 16][@first = "Winston"][@last = "Churchill"]. But implementations must ensure that such rewrites do not result in dynamic errors being reported that would not occur if the predicates were evaluated in order as written.

Note:

Although the rules for guarded expressions prevent optimizations resulting in spurious errors, they do not prevent optimizations whose effect is to mask errors. For example, the rules guarantee that ("A", 3)[. instance of xs:integer][. eq 3] will not raise an error caused by the comparison ("A" eq 3), but they do not guarantee the converse: the expression ("A", 3)[. eq 3][. instance of xs:integer] may or may not raise a dynamic error.

Note:

The rules in this section do not disallow all expression rewrites that might result in dynamic errors. For example, rewriting ($x - $y + $z) as ($x + $z - $y) is permitted even though it might result in an arithmetic overflow.

Note:

Some implementations allow calls on external functions that have side-effects. The semantics of such function calls are entirely implementation defined. Processors may choose to reference the rules for guarded expressions when defining the behavior of such function calls, but this is outside the scope of the language specification.

2.4.6 Implausible Expressions

Changes in 4.0  

  1. The rules for reporting type errors during static analysis have been changed so that a processor has more freedom to report errors in respect of constructs that are evidently wrong, such as @price/@value, even though dynamic evaluation is defined to return an empty sequence rather than an error.   [Issue 602 PR 603 25 July 2023]

[Definition: Certain expressions, while not erroneous, are classified as being implausible, because they achieve no useful effect.]

An example of an implausible expression is @code/text(). This expression will always evaluate to an empty sequence, because attribute nodes cannot have text node children. The semantics of the expression are well defined, but it is likely that the user writing this expression intended something different; if they wanted to an expression that evaluated to an empty sequence, there would be easier ways to write it.

Where an expression is classified (by rules in this specification) as being implausible, a processor may (but is not required to) raise a static error.

For reasons of backwards compatibility and interoperability, and to facilitate automatic generation of XPath 4.0 code, a processor must provide a mode of operation in which implausible expressions are not treated as static errors, but are evaluated with the defined semantics for the expression.

Some other examples of implausible expressions include:

  • round(tokenize($input)). The result of fn:tokenize is a sequence of strings (xs:string*), while the required type for the first argument of fn:round is optional numeric (xs:numeric?). The expression can succeed only in the exceptional case where the result of fn:tokenize is an empty sequence, in which case the result of fn:round will also be an empty sequence; it is therefore highly likely that the expression was written in error.

  • parse-csv($input)?column-names. The signature of the parse-csv function declares its return type as record(columns, rows). There is no field in this record named column-names, and therefore the lookup expression will always return an empty sequence. Again, there is no good reason that a user would write this, so it is likely that it was written in error.

Note:

The specification is deliberately conservative in the choice of constructs that have been classified as implausible. Constructs have not been classified as implausible merely because there are better ways of writing the same thing, but only in cases where it is considered that no user in full understanding of the specification would intentionally write such a construct. All these cases correspond to situations that would be classed as errors in a language with stricter static typing rules.

Note:

In many cases the classification of constructs as implausible is designed to protect users from usability problems that have been found with earlier versions of the language. without introducing backwards incompatibilities.

2.5 Concepts

This section explains some concepts that are important to the processing of XPath 4.0 expressions.

2.5.1 Document Order

An ordering called document order is defined among all the nodes accessible during processing of a given expression, which may consist of one or more trees (documents or fragments). Document order is defined in Section 2.5 Document OrderDM40, and its definition is repeated here for convenience. Document order is a total ordering, although the relative order of some nodes is implementation-dependent. [Definition: Informally, document order is the order in which nodes appear in the XML serialization of a document.] [Definition: Document order is stable, which means that the relative order of two nodes will not change during the processing of a given expression, even if this order is implementation-dependent.] [Definition: The node ordering that is the reverse of document order is called reverse document order.]

Within a tree, document order satisfies the following constraints:

  1. The root node is the first node.

  2. Every node occurs before all of its children and descendants.

  3. Namespace nodes immediately follow the element node with which they are associated. The relative order of namespace nodes is stable but implementation-dependent.

  4. Attribute nodes immediately follow the namespace nodes of the element node with which they are associated. The relative order of attribute nodes is stable but implementation-dependent.

  5. The relative order of siblings is the order in which they occur in the children property of their parent node.

  6. Children and descendants occur before following siblings.

The relative order of nodes in distinct trees is stable but implementation-dependent, subject to the following constraint: If any node in a given tree T1 is before any node in a different tree T2, then all nodes in tree T1 are before all nodes in tree T2.

2.5.2 Typed Value and String Value

Every node has a typed value and a string value, except for nodes whose value is absentDM40. [Definition: The typed value of a node is a sequence of atomic items and can be extracted by applying the Section 2.1.4 fn:dataFO40 function to the node.] [Definition: The string value of a node is a string and can be extracted by applying the Section 2.1.3 fn:stringFO40 function to the node.]

An implementation may store both the typed value and the string value of a node, or it may store only one of these and derive the other as needed. The string value of a node must be a valid lexical representation of the typed value of the node, but the node is not required to preserve the string representation from the original source document. For example, if the typed value of a node is the xs:integer value 30, its string value might be "30" or "0030".

The typed value, string value, and type annotation of a node are closely related. If the node was created by mapping from an Infoset or PSVI, the relationships among these properties are defined by rules in Section 2.8 Schema InformationDM40.

The relationship between typed value and string value for various kinds of nodes is summarized and illustrated by examples below.

  1. For text and document nodes, the typed value of the node is the same as its string value, as an instance of the type xs:untypedAtomic. The string value of a document node is formed by concatenating the string values of all its descendant text nodes, in document order.

  2. The typed value of a comment, namespace, or processing instruction node is the same as its string value. It is an instance of the type xs:string.

  3. The typed value of an attribute node with the type annotation xs:anySimpleType or xs:untypedAtomic is the same as its string value, as an instance of xs:untypedAtomic. The typed value of an attribute node with any other type annotation is derived from its string value and type annotation using the lexical-to-value-space mapping defined in [XML Schema 1.0] or [XML Schema 1.1] Part 2 for the relevant type.

    Example: A1 is an attribute having string value "3.14E-2" and type annotation xs:double. The typed value of A1 is the xs:double value whose lexical representation is 3.14E-2.

    Example: A2 is an attribute with type annotation xs:IDREFS, which is a list datatype whose item type is the atomic datatype xs:IDREF. Its string value is "bar baz faz". The typed value of A2 is a sequence of three atomic items ("bar", "baz"", "faz""), each of type xs:IDREF. The typed value of a node is never treated as an instance of a named list type. Instead, if the type annotation of a node is a list type (such as xs:IDREFS), its typed value is treated as a sequence of the generalized atomic type from which it is derived (such as xs:IDREF).

  4. For an element node, the relationship between typed value and string value depends on the node’s type annotation, as follows:

    1. If the type annotation is xs:untyped or xs:anySimpleType or denotes a complex type with mixed content (including xs:anyType), then the typed value of the node is equal to its string value, as an instance of xs:untypedAtomic. However, if the nilled property of the node is true, then its typed value is the empty sequence.

      Example: E1 is an element node having type annotation xs:untyped and string value "1999-05-31". The typed value of E1 is "1999-05-31", as an instance of xs:untypedAtomic.

      Example: E2 is an element node with the type annotation formula, which is a complex type with mixed content. The content of E2 consists of the character H, a child element named subscript with string value "2", and the character O. The typed value of E2 is "H2O" as an instance of xs:untypedAtomic.

    2. If the type annotation denotes a simple type or a complex type with simple content, then the typed value of the node is derived from its string value and its type annotation in a way that is consistent with schema validation. However, if the nilled property of the node is true, then its typed value is the empty sequence.

      Example: E3 is an element node with the type annotation cost, which is a complex type that has several attributes and a simple content type of xs:decimal. The string value of E3 is "74.95". The typed value of E3 is 74.95, as an instance of xs:decimal.

      Example: E4 is an element node with the type annotation hatsizelist, which is a simple type derived from the atomic type hatsize, which in turn is derived from xs:integer. The string value of E4 is "7 8 9". The typed value of E4 is a sequence of three values (7, 8, 9), each of type hatsize.

      Example: E5 is an element node with the type annotation my:integer-or-string which is a union type with member types xs:integer and xs:string. The string value of E5 is "47". The typed value of E5 is 47 as a xs:integer, since xs:integer is the member type that validated the content of E5. In general, when the type annotation of a node is a union type, the typed value of the node will be an instance of one of the member types of the union.

      Note:

      If an implementation stores only the string value of a node, and the type annotation of the node is a union type, the implementation must be able to deliver the typed value of the node as an instance of the appropriate member type.

    3. If the type annotation denotes a complex type with empty content, then the typed value of the node is the empty sequence and its string value is the zero-length string.

    4. If the type annotation denotes a complex type with element-only content, then the typed value of the node is absentDM40. The fn:data function raises a type error [err:FOTY0012]FO40 when applied to such a node. The string value of such a node is equal to the concatenated string values of all its text node descendants, in document order.

      Example: E6 is an element node with the type annotation weather, which is a complex type whose content type specifies element-only. E6 has two child elements named temperature and precipitation. The typed value of E6 is absentDM40, and the fn:data function applied to E6 raises an error.

2.5.3 Atomization

The semantics of some XPath 4.0 operators depend on a process called atomization. Atomization is applied to a value when the value is used in a context in which a sequence of atomic items is required. The result of atomization is either a sequence of atomic items or a type error [err:FOTY0012]FO40. [Definition: Atomization of a sequence is defined as the result of invoking the fn:data function, as defined in Section 2.1.4 fn:dataFO40.]

The semantics of fn:data are repeated here for convenience. The result of fn:data is the sequence of atomic items produced by applying the following rules to each item in the input sequence:

  • If the item is an atomic item, it is returned.

  • If the item is a node, its typed value is returned (a type error [err:FOTY0012]FO40 is raised if the node has no typed value.)

  • If the item is a function item (other than an array) or map a type error [err:FOTY0013]FO40 is raised.

  • If the item is an array $a, atomization is defined as $a?* ! fn:data(.), which is equivalent to atomizing the members of the array.

    Note:

    This definition recursively atomizes members that are arrays. Hence, the result of atomizing the array [ [ 1, 2, 3 ], [ 4, 5, 6 ] ] is the sequence (1, 2, 3, 4, 5, 6).

Atomization is used in processing the following types of expressions:

  • Arithmetic expressions

  • Comparison expressions

  • Function calls and returns

  • Cast expressions

2.5.4 Effective Boolean Value

Under certain circumstances (listed below), it is necessary to find the effective boolean value of a value. [Definition: The effective boolean value of a value is defined as the result of applying the fn:boolean function to the value, as defined in Section 7.3.1 fn:booleanFO40.]

The dynamic semantics of fn:boolean are repeated here for convenience:

  1. If its operand is an empty sequence, fn:boolean returns false.

  2. If its operand is a sequence whose first item is a node, fn:boolean returns true.

  3. If its operand is a singleton value of type xs:boolean or derived from xs:boolean, fn:boolean returns the value of its operand unchanged.

  4. If its operand is a singleton value of type xs:string, xs:anyURI, xs:untypedAtomic, or a type derived from one of these, fn:boolean returns false if the operand value has zero length; otherwise it returns true.

  5. If its operand is a singleton value of any numeric type or derived from a numeric type, fn:boolean returns false if the operand value is NaN or is numerically equal to zero; otherwise it returns true.

  6. In all other cases, fn:boolean raises a type error [err:FORG0006]FO40.

    Note:

    For instance, fn:boolean raises a type error if the operand is a function, a map, or an array.

The effective boolean value of a sequence is computed implicitly during processing of the following types of expressions:

  • Logical expressions (and, or)

  • The fn:not function

  • Certain types of predicates, such as a[b]

  • Conditional expressions (if)

  • Quantified expressions (some, every)

  • General comparisons, in XPath 1.0 compatibility mode.

Note:

The definition of effective boolean value is not used when casting a value to the type xs:boolean, for example in a cast expression or when passing a value to a function whose expected parameter is of type xs:boolean.

2.5.5 URI Literals

XPath 4.0 requires a statically known, valid URI in a BracedURILiteral. An implementation may raise a static error [err:XQST0046] if the value of a Braced URI Literal is of nonzero length and is neither an absolute URI nor a relative URI.

Note:

The xs:anyURI type is designed to anticipate the introduction of Internationalized Resource Identifiers (IRIs) as defined in [RFC3987].

Whitespace is normalized using the whitespace normalization rules of fn:normalize-space. If the result of whitespace normalization contains only whitespace, the corresponding URI consists of the empty string.

A Braced URI Literal or URI Literal is not subjected to percent-encoding or decoding as defined in [RFC3986].

2.5.6 Resolving a Relative URI Reference

[Definition: To resolve a relative URI $rel against a base URI $base is to expand it to an absolute URI, as if by calling the function fn:resolve-uri($rel, $base).] During static analysis, the base URI is the Static Base URI. During dynamic evaluation, the base URI used to resolve a relative URI reference depends on the semantics of the expression.

Any process that attempts to resolve a URI against a base URI, or to dereference the URI, may apply percent-encoding or decoding as defined in the relevant RFCs.

3 Types

As noted in 2.1.2 Values, every value in XPath 4.0 is regarded as a sequence of zero, one, or more items. The type system of XPath 4.0, described in this section, classifies the kinds of value that the language can handle, and the operations permitted on different kinds of value.

The type system of XPath 4.0 is related to the type system of [XML Schema 1.0] or [XML Schema 1.1] in two ways:

This chapter of the specification starts by defining sequence types and item types, which describe the range of values that can be bound to variables, used in expressions, or passed to functions. It then describes how these relate to schema types, that is, the simple and complex types defined in an XSD schema.

3.1 Sequence Types

[Definition: A sequence type is a type that can be expressed using the SequenceType syntax. Sequence types are used whenever it is necessary to refer to a type in an XPath 4.0 expression. The term sequence type suggests that this syntax is used to describe the type of an XPath 4.0 value, which is always a sequence.]

[Definition: An item type is a type that can be expressed using the ItemType syntax, which forms part of the SequenceType syntax. Item types match individual items.] In most cases, the set of items matched by an item type consists either exclusively of atomic items, exclusively of nodes, or exclusively of function itemsDM40. Exceptions include the generic types item(), which matches all items, xs:error, which matches no items, and choice item types, which can match any combination of types.

Whenever it is necessary to refer to a type in an XPath 4.0 expression, the SequenceType syntax is used.

SequenceType ::= ("empty-sequence" "(" ")")
| (ItemType OccurrenceIndicator?)
ItemType ::= AnyItemTest | TypeName | KindTest | FunctionType | MapType | ArrayType | RecordType | EnumerationType | ChoiceItemType
OccurrenceIndicator ::= "?" | "*" | "+"
/* xgc: occurrence-indicators */

In many situations the terms item type and sequence type are used interchangeably to refer either to the type itself, or to the syntactic construct that designates the type: so in the expression $x instance of xs:string*, the construct xs:string* uses the SequenceType syntax to designate a sequence type whose instances are sequences of strings. When more precision is required, the specification is careful to use the terms item type and sequence type to refer to the actual types, while using the production names ItemType and SequenceType to refer to the syntactic designators of these types.

[Definition: A sequence type designator is a syntactic construct conforming to the grammar rule SequenceType. A sequence type designator is said to designate a sequence type.]

[Definition: An item type designator is a syntactic construct conforming to the grammar rule ItemType. An item type designator is said to designate an item type.]

Note:

Two item type designators may designate the same item type. For example, element() and element(*) are equivalent, as are attribute(A) and attribute(A, xs:anySimpleType).

With the exception of the special type empty-sequence(), a sequence type consists of an item type that constrains the type of each item in the sequence, and a cardinality that constrains the number of items in the sequence. Apart from the item type item(), which permits any kind of item, item types divide into node types (such as element()), generalized atomic types (such as xs:integer) and function types (such as function() as item()*).

Lexical QNames appearing in a sequence type have their prefixes expanded to namespace URIs by means of the statically known namespaces and (where applicable) the default namespace for elements and types. Equality of QNames is defined by the eq operator.

Item types representing element and attribute nodes may specify the required type annotations of those nodes in the form of a schema type. Thus the item type element(*, us:address) denotes any element node whose type annotation is (or is derived from) the schema type named us:address.

The occurrence indicators +, *, and ? bind to the last ItemType in the SequenceType, as described in the occurrence-indicators constraint.

3.1.1 Examples of Sequence Types

Here are some examples of sequence types that might be used in XPath 4.0:

  • xs:date refers to the built-in atomic schema type named xs:date

  • attribute()? refers to an optional attribute node

  • element() refers to any element node

  • element(po:shipto, po:address) refers to an element node that has the name po:shipto and has the type annotation po:address (or a schema type derived from po:address)

  • element(*, po:address) refers to an element node of any name that has the type annotation po:address (or a type derived from po:address)

  • element(customer) refers to an element node named customer with any type annotation

  • schema-element(customer) refers to an element node whose name is customer (or is in the substitution group headed by customer) and whose type annotation matches the schema type declared for a customer element in the in-scope element declarations

  • node()* refers to a sequence of zero or more nodes of any kind

  • item()+ refers to a sequence of one or more items

  • function(*) refers to any function item, regardless of arity or type

  • function(node()) as xs:string* refers to a function item that takes a single argument whose value is a single node, and returns a sequence of zero or more xs:string values

  • (fn(node()) as xs:string)* refers to a sequence of zero or more function items, each of which takes a single argument whose value is a single node, and returns as its result a single xs:string value

3.1.2 Sequence Type Matching

[Definition: SequenceType matching compares a value with an expected sequence type. ] For example, an instance of expression returns true if a given value matches a given sequence type, and false if it does not.

An XPath 4.0 implementation must be able to determine relationships among the types in type annotations in an XDM instance and the types in the in-scope schema definitions (ISSD).

[Definition: The use of a value that has a dynamic type that is a subtype of the expected type is known as subtype substitution.] Subtype substitution does not change the actual type of a value. For example, if an xs:integer value is used where an xs:decimal value is expected, the value retains its type as xs:integer.

The definition of SequenceType matching relies on a pseudo-function named derives-from( AT, ET ), which takes an actual simple or complex schema type AT and an expected simple or complex schema type ET, and either returns a boolean value or raises a type error [err:XPTY0004]. This function is defined as follows:

  • derives-from( AT, ET ) raises a type error [err:XPTY0004] if ET is not present in the in-scope schema definitions (ISSD).

  • derives-from( AT, ET ) returns true if any of the following conditions applies:

    • AT is ET

    • ET is the base type of AT

    • ET is a pure union type of which AT is a member type

    • There is a type MT such that derives-from( AT, MT ) and derives-from( MT, ET )

  • Otherwise, derives-from( AT, ET ) returns false

The rules for SequenceType matching are given below, with examples (the examples are for purposes of illustration, and do not cover all possible cases).

An OccurrenceIndicator specifies the number of items in a sequence, as follows:

  • ? matches zero or one items

  • * matches zero or more items

  • + matches one or more items

As a consequence of these rules, any sequence type whose OccurrenceIndicator is * or ? matches a value that is an empty sequence.

3.2 Item Types

ItemType ::= AnyItemTest | TypeName | KindTest | FunctionType | MapType | ArrayType | RecordType | EnumerationType | ChoiceItemType
AnyItemTest ::= "item" "(" ")"
TypeName ::= EQName
KindTest ::= DocumentTest
| ElementTest
| AttributeTest
| SchemaElementTest
| SchemaAttributeTest
| PITest
| CommentTest
| TextTest
| NamespaceNodeTest
| AnyKindTest
DocumentTest ::= "document-node" "(" (ElementTest | SchemaElementTest | NameTestUnion)? ")"
ElementTest ::= "element" "(" (NameTestUnion ("," TypeName "?"?)?)? ")"
SchemaElementTest ::= "schema-element" "(" ElementName ")"
AttributeTest ::= "attribute" "(" (NameTestUnion ("," TypeName)?)? ")"
SchemaAttributeTest ::= "schema-attribute" "(" AttributeName ")"
ElementName ::= EQName
AttributeName ::= EQName
PITest ::= "processing-instruction" "(" (NCName | StringLiteral)? ")"
CommentTest ::= "comment" "(" ")"
NamespaceNodeTest ::= "namespace-node" "(" ")"
TextTest ::= "text" "(" ")"
AnyKindTest ::= "node" "(" ")"
FunctionType ::= AnyFunctionType
| TypedFunctionType
AnyFunctionType ::= ("function" | "fn") "(" "*" ")"
TypedFunctionType ::= ("function" | "fn") "(" (SequenceType ** ",") ")" "as" SequenceType
ChoiceItemType ::= "(" (ItemType ++ "|") ")"
MapType ::= AnyMapType | TypedMapType
RecordType ::= AnyRecordType | TypedRecordType
ArrayType ::= AnyArrayType | TypedArrayType
EnumerationType ::= "enum" "(" (StringLiteral ++ ",") ")"

This section defines the syntax and semantics of different ItemTypes in terms of the values that they match.

Note:

For an explanation of the EBNF grammar notation (and in particular, the operators ++ and **), see A.1 EBNF.

An item type designator written simply as an EQName (that is, a TypeName) is interpreted as follows:

  1. If the name is written as a lexical QName, then it is expanded using the in-scope namespaces in the static context. If the name is an unprefixed NCName, then it is expanded according to the default namespace for elements and types.

  2. If the name matches a named item type in the static context, then it is taken as a reference to the corresponding item type. The rules that apply are the rules for the expanded item type definition.

  3. Otherwise, it must match the name of a type in the in-scope schema types in the static context: specifically, an atomic type or a pure union type. See 3.5 Schema Types for details.

    Note:

    A name in the xs namespace will always fall into this category, since the namespace is reserved. See 2.1.3 Namespaces and QNames.

  4. If the name cannot be resolved to a type, a static error is raised [err:XPST0051].

3.2.1 General item types

  • item() matches any single item.

    For example, item() matches the atomic item 1, the element <a/>, or the function fn:concat#3.

  • A ChoiceItemType lists a number of alternative item types in parentheses, separated by "|". An item matches a ChoiceItemType it if matches any of the alternatives.

    For example, (map(*) | array(*)) matches any item that is a map or an array.

    Note:

    If there is only one alternative, the ChoiceItemType designates the same item type as the ItemType that is in parentheses. A singleton choice (that is, a parenthesized item type) is used primarily when defining nested item types in a function signature. For example, a sequence of functions that each return a single boolean might be denoted (fn() as xs:boolean)*. In this example the parentheses are needed to indicate where the occurrence indicator belongs.

3.2.2 Atomic Types

Atomic types in the XPath 4.0 type system correspond directly to atomic types as defined in the [XML Schema 1.0] or [XML Schema 1.1] type system.

Atomic types are either built-in atomic types such as xs:integer, or user-defined atomic types imported from a schema. Atomic types are identified by a QName: see 2.1.3 Namespaces and QNames.

Note:

A schema may also include anonymous atomic types. Such types are not usable directly in XPath 4.0, though they may appear as the values of type annotations on nodes.

[Definition: A generalized atomic type is an item type whose instances are all atomic items. Generalized atomic types include (a) atomic types, either built-in (for example xs:integer) or imported from a schema, (b) pure union types, either built-in (xs:numeric and xs:error) or imported from a schema, (c) choice item types if their alternatives are all generalized atomic types, and (d) enumeration types. ].

A generalized atomic type may be designated by an ItemType in any of the following ways:

An atomic item A matches the generalized atomic type GAT if the type annotation of A (call it T) satisfies the condition derives-from(T, GAT).

Example: The ItemType xs:decimal matches any value of type xs:decimal. It also matches any value of type shoesize, if shoesize is an atomic type derived by restriction from xs:decimal.

Example: Suppose ItemType dress-size is a union type that allows either xs:decimal values for numeric sizes (for example: 4, 6, 10, 12), or one of an enumerated set of xs:strings (for example: small, medium, large). The ItemType dress-size matches any of these values.

Note:

The names of list types such as xs:IDREFS are not accepted in this context, but can often be replaced by a generalized atomic type with an occurrence indicator, such as xs:IDREF+.

3.2.3 Union Types

Union types, as defined in XSD, are a variety of simple types. The membership of a union type in XSD may include list types as well as atomic types and other union types.

[Definition: A pure union type is a simple type that satisfies the following constraints: (a) {variety} is union, (b) the {facets} property is empty, (c) no type in the transitive membership of the union type has {variety} list, and (d) no type in the transitive membership of the union type is a type with {variety} union having a non-empty {facets} property].

Note:

The definition of pure union type excludes union types derived by non-trivial restriction from other union types, as well as union types that include list types in their membership. Pure union types have the property that every instance of an atomic type defined as one of the member types of the union is also a valid instance of the union type.

Note:

The current (second) edition of XML Schema 1.0 contains an error in respect of the substitutability of a union type by one of its members: it fails to recognize that this is unsafe if the union is derived by restriction from another union.

This problem is fixed in XSD 1.1, but the effect of the resolution is that an atomic item labeled with an atomic type cannot be treated as being substitutable for a union type without explicit validation. This specification therefore allows union types to be used as item types only if they are defined directly as the union of a number of atomic types.

Note:

Local union types (see 3.2.5 Choice Item Types) and enumeration types cannot be used as the target for schema validation.

3.2.4 Namespace-sensitive Types

[Definition: The namespace-sensitive types are xs:QName, xs:NOTATION, types derived by restriction from xs:QName or xs:NOTATION, list types that have a namespace-sensitive item type, and union types with a namespace-sensitive type in their transitive membership.]

It is not possible to preserve the type of a namespace-sensitive value without also preserving the namespace binding that defines the meaning of each namespace prefix used in the value. Therefore, XPath 4.0 defines some error conditions that occur only with namespace-sensitive values. For instance, casting to a namespace-sensitive type raises a type error [err:FONS0004]FO40 if the namespace bindings for the result cannot be determined.

3.2.5 Choice Item Types

Changes in 4.0  

  1. Choice item types (an item type allowing a set of alternative item types) are introduced.   [Issue 122 PR 1132 9 April 2024]

[Definition: A choice item type defines an item type that is the union of a number of alternatives. For example the type (xs:hexBinary | xs:base64Binary) defines the union of these two primitive atomic types, while the type (map(*) | array(*)) matches any item that is either a map or an array.]

An item matches a ChoiceItemType if it matches any of the alternatives listed within the parentheses.

For example, the type (xs:NCName | enum("")) matches any value that is either an instance of xs:NCName, or a zero-length string. This might be a suitable type for a variable that holds a namespace prefix.

If all the alternatives are generalized atomic types then the choice item type is itself a generalized atomic type, which means, for example, that it can be used as the target of a cast expression.

Note:

A choice item type in which all the alternatives are atomic behaves in most respects like a schema-defined pure union type. However, because it can be defined at the point of use (for example, within a function signature), it may be more convenient than defining the type in an imported schema.

Note:

Choice item types are particularly useful in function signatures, allowing a function to take arguments of a variety of types. If the choice item type is a local union type, then the semantics are identical to using a named union type, but a local union type is more convenient because it does not need to be defined in a schema, and does not require a schema-aware processor.

A local union type can also be used in a cast expression: cast @when as (xs:date | xs:dateTime) allows the attribute @when to be either an xs:date, or an xs:dateTime.

An instance of expression can be used to test whether a value belongs to one of a number of specified types: $x instance of (xs:string | xs:anyURI | xs:untypedAtomic) returns true if $x is an instance of any of these three atomic types, while $x instance of (map(*) | array(*)) tests whether $x is a map or array.

3.2.6 Enumeration Types

Changes in 4.0  

  1. Enumeration types are added as a new kind of ItemType, constraining the value space of strings.  [Issue 688 PR 691 10 October 2023]

[Definition: An EnumerationType accepts a fixed set of string values.]

EnumerationType ::= "enum" "(" (StringLiteral ++ ",") ")"

An enumeration type has a value space consisting of a set of xs:string values. When matching strings against an enumeration type, strings are always compared using the Unicode codepoint collation.

For example, if an argument of a function declares the required type as enum("red", "green", "blue"), then the string "green" is accepted, while "yellow" is rejected with a type error.

Technically, enumeration types are defined as follows:

  • An enumeration type with a single enumerated value (such as enum("red")) is an anonymous atomic type derived from xs:string by restriction using an enumeration facet that permits only the value "red". This is referred to as a singleton enumeration type. It is equivalent to the XSD-defined type:

    <xs:simpleType>
      <xs:restriction base="xs:string">
        <xs:enumeration value="red"/>
      </xs:restriction>
    </xs:simpleType>
  • Two singleton enumeration types are the same type if and only if they have the same (single) enumerated value, as determined using the Unicode codepoint collation.

  • An enumeration type with multiple enumerated values is a union of singleton enumeration types, so enum("red", "green", "blue") is equivalent to (enum("red") | enum("green") | enum("blue")).

  • In consequence, an enumeration type T is a subtype of an enumeration type U if the enumerated values of T are a subset of the enumerated values of U: see 3.3.2 Subtypes of Item Types.

An enumeration type is thus a generalized atomic type.

It follows from these rules that an atomic item will only satisfy an instance of test if it has the correct type annotation, and this can only be achieved using an explicit cast or constructor function. So the expression "red" instance of enum("red", "green", "blue") returns false. However, the coercion rules ensure that where a variable or function declaration specifies an enumeration type as the required type, a string (or indeed an xs:untypedAtomic or xs:anyURI value) equal to one of the enumerated values will be accepted.

Note:

Some consequences of these rules may not be immediately apparent.

Suppose that an XQuery query contains the declarations:

declare type my:color := enum("red", "green", "orange");
declare type my:fruit := enum("apple", "orange", "banana");
declare variable $orange-color as my:color := "orange";
declare variable $orange-fruit as my:fruit := "orange";

The same applies with the equivalent XSLT syntax:

<xsl:item-type name="my:color" as="enum('red', 'green', 'orange')"/>
<xsl:item-type name="my:fruit" as="enum('apple', 'orange', 'banana')"/>
<xsl:variable name="orange-color" as="my:color" select="'orange'"/>
<xsl:variable name="orange-fruit" as="my:fruit" select="'orange'"/>

Now, the value of $orange-color is an atomic item whose datum is the string "orange", and whose type annotation is the anonymous type designated enum("orange"). Similarly, the value of $orange-fruit is an atomic item whose datum is the string "orange", and whose type annotation is the anonymous type designated enum("orange"). That is, the values of the two variables are indistinguishable and interchangeable in every way. In particular, both values are instances of my:color, and both are instances of my:fruit.

This way of handling enumeration values has advantages and disadvantages. On the positive side, it means that enumeration subsets and supersets work cleanly: a value that is an instance of enum("red", "green", "orange") can be used where an instance of enum("red", "orange", "yellow", "green", "blue", "indigo", "violet") is expected. The downside is that labeling a string as an instance of an enumeration type does not provide type safety: a function that expects an instance of my:color can be called with any string that matches one of the required colors, whether or not it has an appropriate type annotation. A function that expects a color can be successfully called passing a fruit, if they happen to have the same name.

3.2.7 Node Types

Changes in 4.0  

  1. Element and attribute tests can include alternative names: element(chapter|section), attribute(role|class).   [Issue 107 PR 286 17 January 2023]

  2. The NodeTest in an AxisStep now allows alternatives: ancestor::(section|appendix)   [Issue 107 PR 286 17 January 2023]

  3. The syntax document-node(N), where N is a NameTestUnion, is introduced as an abbreviation for document-node(element(N)). For example, document-node(*) matches any well-formed XML document (as distinct from a document fragment).   [Issue 1593  24 November 2024]

Some of the constructs described in this section include a TypeName. This appears as T in:

  • element(N, T)

  • attribute(N, T)

  • document-node(element(N, T))

In these constructs, the type name T is expanded using the in-scope namespaces in the static context, using the default namespace for elements and types if it is unprefixed. The resulting QName must identify a type in the in-scope schema definitions. This can be any schema type: either a simple type, or (except in the case of attributes) a complex type. If it is a simple type then it can be an atomic, union, or list type. It can be a built-in type (such as xs:integer) or a user-defined type. It must however be the name of a type defined in a schema; it cannot be a named item type.

3.2.7.1 Simple Node Tests
  • node() matches any node.

  • text() matches any text node.

  • processing-instruction() matches any processing-instruction node.

  • processing-instruction( N ) matches any processing-instruction node whose PITarget is equal to fn:normalize-space(N). If the result of fn:normalize-space(N) is not in the lexical space of NCName, a type error is raised [err:XPTY0004]

    Example: processing-instruction(xml-stylesheet) matches any processing instruction whose PITarget is xml-stylesheet.

    For backward compatibility with XPath 1.0, the PITarget of a processing instruction may also be expressed as a string literal, as in this example: processing-instruction("xml-stylesheet").

    If the specified PITarget is not a syntactically valid NCName, a type error is raised [err:XPTY0004].

  • comment() matches any comment node.

  • namespace-node() matches any namespace node.

  • document-node() matches any document node.

  • document-node( E ), where E is an ElementTest or SchemaElementTest (see 3.2.7.2 Element Test and 3.2.7.3 Schema Element Test) matches any document node whose children comprise (in any order) zero or more comment and processing instruction nodes, zero text nodes, and exactly one element node, which itself must match E.

    Example: document-node(element(book)) matches a document node containing exactly one element node that is matched by the ElementTest element(book).

  • The construct document-node(NTU), where NTU is a NameTestUnion, is an abbreviation for document-node(element(NTU)). For example, document-node(*) is an abbreviation for document-node(element(*)), which matches any document node corresponding to a well-formed XML document, that is, one that has one element child, zero or more comment and processing-instruction children, and no text node children.

  • An ItemType that is an ElementTest, SchemaElementTest, AttributeTest, SchemaAttributeTest, or FunctionType matches an item as described in the following sections.

3.2.7.2 Element Test

Changes in 4.0  

  1. Element and attribute tests of the form element(N) and attribute(N) now allow N to be any NameTest, including a wildcard. The forms element(A|B) and attribute(A|B) are also allowed.   [Issue 107 PR 286 17 January 2023]

  2. Setting the default namespace for elements and types to the special value ##any causes an unprefixed element name to act as a wildcard, matching by local name regardless of namespace.

ElementTest ::= "element" "(" (NameTestUnion ("," TypeName "?"?)?)? ")"
NameTestUnion ::= (NameTest ++ "|")
NameTest ::= EQName | Wildcard
Wildcard ::= "*"
| (NCName ":*")
| ("*:" NCName)
| (BracedURILiteral "*")
/* ws: explicit */
TypeName ::= EQName

An ElementTest is used to match an element node by its name and/or type annotation.

An unprefixed EQName within the NameTestUnion is interpreted according to the default namespace for elements and types. The name need not be present in the in-scope element declarations.

If the default namespace for elements and types has the special value ##any, then an unprefixed name N is interpreted as a wildcard *:N .

It is always possible to match no-namespace names explicitly by using the form Q{}N

An unprefixed TypeName is interpreted according to the default namespace for elements and types. The TypeName must be present in the in-scope schema types [err:XPST0008]

If the default namespace for elements and types has the special value ##any, then an unprefixed type name T is interpreted as Q{http://www.w3.org/2001/XMLSchema}T .

Note:

Substitution groups do not affect the semantics of ElementTest.

An ElementTest ET matches an item E if the following conditions are satisfied:

  1. E is an element node.

  2. If ET includes a NameTestUnion, then the name of the element node E matches one or more of the NameTests in the NameTestUnion. A name N matches a NameTest NT if one of the following conditions is true:

    1. NT is *

    2. NT is *:local and the local part of N is local.

    3. NT is prefix:* and the namespace URI of N matches the namespace URI bound to prefix in the static context.

    4. NT is BracedURILiteral* and the namespace URI of N matches the namespace URI found in the BracedURILiteral.

    5. NT is an EQName equal to N.

  3. If ET includes a TypeName, then the type annotation of the element node E is either the schema type identified by that type name, or a type derived from that type by restriction.

  4. If E has the nilled property, then ET either includes no TypeName, or includes a TypeName followed by the symbol ?.

Here are some examples of ElementTests:

  1. element() and element(*) match any single element node, regardless of its name or type annotation.

  2. element(person) matches any element node whose name is person, in the default namespace for elements and types.

  3. element(doctor|nurse) matches any element node whose name is doctor or nurse, in the default namespace for elements and types.

  4. element(xhtml:*) matches any element node whose name is in the namespace bound to the prefix xhtml.

  5. element(xhtml:*|svg:*|mathml:*) matches any element node whose name is one of the three namespaces identified, specifically the namespaces bound to the prefixes xhtml, svg, and mathml.

  6. element(Q{http://www.w3.org/2000/svg}*) matches any element node whose name is in the SVG namespace.

  7. element(*:html) matches any element node whose local name is "html", in any namespace.

  8. element(person, surgeon) matches a non-nilled element node whose name is person and whose type annotation is surgeon (or is derived from surgeon).

  9. element(person, surgeon?) matches a nilled or non-nilled element node whose name is person and whose type annotation is surgeon (or is derived from surgeon).

  10. element(*, surgeon) matches any non-nilled element node whose type annotation is surgeon (or is derived from surgeon), regardless of its name.

  11. element(*, surgeon?) matches any nilled or non-nilled element node whose type annotation is surgeon (or is derived from surgeon), regardless of its name.

Where a TypeName is included in an ElementTest T, and element node will only match the test if it has been validated against a schema that defines type T; furthermore, T must be present in the in-scope schema definitions of the static context of the ElementTest. Although it is guaranteed that type T will have compatibleDM40 definitions in the schema that was used for validation and in the in-scope schema definitions, it is not guaranteed that revalidation using the in-scope schema definitions would succeed. For example, if substitution group membership varies between the two schemas, the element node may contain children or descendants that the in-scope schema definitions would not allow.

Note:

Technically, element(p|q) is not the same type as the choice item type (element(p)|element(q)). However, (a) they match exactly the same set of element nodes, and (b) each is a subtype of the other, so in practice they are indistinguishable.

3.2.7.3 Schema Element Test
SchemaElementTest ::= "schema-element" "(" ElementName ")"
ElementName ::= EQName

A SchemaElementTest matches an element node against a corresponding element declaration found in the in-scope element declarations.

The ElementName of a SchemaElementTest has its prefixes expanded to a namespace URI by means of the statically known namespaces, or if unprefixed, the is interpreted according to the default namespace for elements and types. If this has the special value "##any", an unprefixed name represents a name in no namespace. If the ElementName specified in the SchemaElementTest is not found in the in-scope element declarations, a static error is raised [err:XPST0008].

A SchemaElementTest matches a candidate element node if all of the following conditions are satisfied:

  1. Either:

    1. The name N of the candidate node matches the specified ElementName, or

    2. The name N of the candidate node matches the name of an element declaration that is a member of the actual substitution group headed by the declaration of element ElementName.

    Note:

    The term “actual substitution group” is defined in [XML Schema 1.1]. The actual substitution group of an element declaration H includes those element declarations P that are declared to have H as their direct or indirect substitution group head, provided that P is not declared as abstract, and that P is validly substitutable for H, which means that there must be no blocking constraints that prevent substitution.

  2. The schema element declaration named N is not abstract.

  3. derives-from( AT, ET ) is true, where AT is the type annotation of the candidate node and ET is the schema type declared in the schema element declaration named N.

  4. If the schema element declaration named N is not nillable, then the nilled property of the candidate node is false.

Example: The SchemaElementTest schema-element(customer) matches a candidate element node in the following two situations:

  1. customer is a top-level element declaration in the in-scope element declarations; the name of the candidate node is customer; the element declaration of customer is not abstract; the type annotation of the candidate node is the same as or derived from the schema type declared in the customer element declaration; and either the candidate node is not nilled, or customer is declared to be nillable.

  2. customer is a top-level element declaration in the in-scope element declarations; the name of the candidate node is client; client is an actual (non-abstract and non-blocked) member of the substitution group of customer; the type annotation of the candidate node is the same as or derived from the schema type declared for the client element; and either the candidate node is not nilled, or client is declared to be nillable.

In the case where the schema X used to validate an element node E (whose name is N) differs from the schema Y represented by the in-scope schema definitions in the static context of the SchemaElementTest, the following considerations apply:

  • In applying the test derives-from( AT, ET ), note that AT will necessarily be present in X, but not necessarily in Y. However, ET will necessarily be present in both; and because the two schemas must be compatibleDM40, ET will be the present in both schemas, will have the same definition in both, and will be the declared type of N in both. The test can therefore be applied from knowledge of type AT as defined in schema X.

  • The test as to whether the element name N is a member of the actual substitution group is performed entirely by reference to schema Y. Although the two schemas are compatible, substitution group membership can vary.

3.2.7.4 Attribute Test

Changes in 4.0  

  1. Element and attribute tests of the form element(N) and attribute(N) now allow N to be any NameTest, including a wildcard. The forms element(A|B) and attribute(A|B) are also allowed.   [Issue 107 PR 286 17 January 2023]

AttributeTest ::= "attribute" "(" (NameTestUnion ("," TypeName)?)? ")"
NameTestUnion ::= (NameTest ++ "|")
NameTest ::= EQName | Wildcard
Wildcard ::= "*"
| (NCName ":*")
| ("*:" NCName)
| (BracedURILiteral "*")
/* ws: explicit */
TypeName ::= EQName

An AttributeTest is used to match an attribute node by its name and/or type annotation.

An unprefixed EQName within the NameTestUnion refers to a name in no namespace. The name need not be present in the in-scope attribute declarations.

An unprefixed TypeName is interpreted according to the default namespace for elements and types. The TypeName must be present in the in-scope schema types [err:XPST0008]

An AttributeTest AT matches an item A if the following conditions are satisfied:

  1. A is an attribute node.

  2. If AT includes a NameTestUnion, then the name of the attribute node A matches one or more of the NameTests in the NameTestUnion. A name N matches a NameTest NT if one of the following conditions is true:

    1. NT is *

    2. NT is *:local and the local part of N matches local.

    3. NT is prefix:* and the namespace URI of N matches the namespace URI bound to prefix in the static context.

    4. NT is BracedURILiteral* and the namespace URI of N matches the namespace URI found in the BracedURILiteral.

    5. NT is an EQName equal to N.

  3. If AT includes a TypeName, then the type annotation of the attribute node A is either the schema type identified by that type name, or a type derived from that type by restriction.

Here are some examples of AttributeTests:

  • attribute() and attribute(*) match any single attribute node, regardless of its name or type annotation.

  • attribute(price) matches any attribute node whose name is price (in no namespace), regardless of its type annotation.

  • attribute(price|discount) matches any attribute node whose name is price or discount (in no namespace).

  • attribute(xlink:*) matches any attribute node whose name is in the namespace bound to the prefix xlink.

  • element(Q{http://www.w3.org/2000/svg}*) matches any attribute node whose name is in the SVG namespace.

  • attribute(*:default-collation) matches any attribute node whose local name is default-collation, regardless of namespace, and regardless of type annotation.

  • attribute(*:price|*:discount) matches any attribute node whose local name is price or discount, regardless of namespace, and regardless of type annotation.

  • attribute(price, currency) matches an attribute node whose name is price (in no namespace) and whose type annotation is currency (or is derived from currency).

  • attribute(xlink:*, xs:string) matches any attribute node whose name is in the namespace bound to the prefix xlink, and whose type annotation is xs:string or a type derived from xs:string.

  • attribute(*, currency) matches any attribute node whose type annotation is currency (or is derived from currency), regardless of its name.

Unlike the situation with an ElementTest, few problems arise if the attribute was validated using a different schema. This is because simple types can never be derived by extension, and attributes do not have substitution groups.

Note:

Technically, attribute(p|q) is not the same type as the choice item type (attribute(p)|attribute(q)). However, (a) they match exactly the same set of attribute nodes, and (b) each is a subtype of the other, so in practice they are indistinguishable.

3.2.7.5 Schema Attribute Test
SchemaAttributeTest ::= "schema-attribute" "(" AttributeName ")"
AttributeName ::= EQName

A SchemaAttributeTest matches an attribute node against a corresponding attribute declaration found in the in-scope attribute declarations.

The AttributeName of a SchemaAttributeTest has its prefixes expanded to a namespace URI by means of the statically known namespaces. If unprefixed, an AttributeName is in no namespace. If the AttributeName specified in the SchemaAttributeTest is not found in the in-scope attribute declarations, a static error is raised [err:XPST0008].

A SchemaAttributeTest matches a candidate attribute node if both of the following conditions are satisfied:

  1. The name of the candidate node matches the specified AttributeName.

  2. derives-from( AT, ET ) is true, where AT is the type annotation of the candidate node and ET is the schema type declared for attribute AttributeName in the in-scope attribute declarations.

Example: The SchemaAttributeTest schema-attribute(color) matches a candidate attribute node if color is a top-level attribute declaration in the in-scope attribute declarations, the name of the candidate node is color, and the type annotation of the candidate node is the same as or derived from the schema type declared for the color attribute.

Unlike the situation with a SchemaElementTest, few problems arise if the attribute was validated using a different schema. This is because attributes do not have substitution groups.

3.2.8 Function, Map, and Array Types

The following sections describe the syntax for item types for functions, including arrays and maps.

The subtype relation among these types is described in the various subsections of 3.3.2 Subtypes of Item Types.

3.2.8.1 Function Type

Changes in 4.0  

  1. The keyword fn is allowed as a synonym for function in function types, to align with changes to inline function declarations.  [Issue 1192 PR 1197 21 May 2024]

  2. The terms FunctionType, ArrayType, MapType, and RecordType replace FunctionTest, ArrayTest, MapTest, and RecordTest, with no change in meaning.

FunctionType ::= AnyFunctionType
| TypedFunctionType
AnyFunctionType ::= ("function" | "fn") "(" "*" ")"
TypedFunctionType ::= ("function" | "fn") "(" (SequenceType ** ",") ")" "as" SequenceType

A FunctionType matches selected function items, potentially checking their signatureDM40 (which includes the types of the arguments and results).

An AnyFunctionType matches any item that is a function.

A TypedFunctionType matches an item if it is a function item and the function’s type signature (as defined in Section 2.9.4 Function ItemsDM40) is a subtype of the TypedFunctionType.

Note:

The keywords function and fn are synonymous.

In addition, a TypedFunctionType may match certain maps and arrays, as described in 3.2.8.2 Map Type and 3.2.8.5 Array Type

Here are some examples of FunctionTypes:

  1. function(*) matches any function, including maps and arrays.

    Note:

    This can also be written fn(*).

  2. function(xs:int, xs:int) as xs:int matches any function item with the function signature function(xs:int, xs:int) as xs:int.

    Note:

    This can also be written fn(xs:int, xs:int) as xs:int.

  3. function(xs:anyAtomicType) as item()* matches any map, or any function with the required signature.

  4. function(xs:integer) as item()* matches any array, or any function with the required signature.

3.2.8.2 Map Type
MapType ::= AnyMapType | TypedMapType
AnyMapType ::= "map" "(" "*" ")"
TypedMapType ::= "map" "(" ItemType "," SequenceType ")"

The MapType map(*) matches any map. The MapType map(K, V) matches any map where every key is an instance of K and every value is an instance of V.

Although the grammar for TypedMapType allows the key to be described using the full ItemType syntax, the item type used must be a generalized atomic type. [TODO: error code].

Examples:

Given a map $M whose keys are integers and whose results are strings, such as { 0: "no", 1: "yes" }, consider the results of the following expressions:

  • $M instance of map(*) returns true()

  • $M instance of map(xs:integer, xs:string) returns true()

  • $M instance of map(xs:decimal, xs:anyAtomicType) returns true()

  • not($M instance of map(xs:int, xs:string)) returns true()

  • not($M instance of map(xs:integer, xs:token)) returns true()

A map is also a function item, and therefore matches certain function types. Specifically, a map that matches map(K, V) also matches a function type of the form function(xs:anyAtomicType) as R provided that both the following conditions are satisfied:

Note:

To understand this rule, consider the use of a map $M in a function call $M($K), which is equivalent to the function call map:get($M, $K). This function accepts any atomic item for the argument $K, and hence matches a function type that requires an argument type of xs:anyAtomicType. If the key $K is present in the map, the result of the function will be a value of type V; if not, it will be an empty sequence. The map is therefore substitutable for the function type provided that the function type allows both a value of type V and the empty sequence as possible results.

The key type K does not enter into this rule. That is because in the function call $M($K), the sought key $K does not have to be of the same type as the keys actually present in the map.

The transitivity rules for item type matching mean that if an item M matches a type T, and T is a subtype of U, then M also matches type U. So the fact that a map from integers to strings (map(xs:integer, xs:string)) matches function(xs:anyAtomicType) as xs:string? means that it will also match other function types such as function(xs:integer) as xs:string? and function(xs:decimal) as xs:anyAtomicType?

Furthermore, the rules for function coercion mean that any map can be supplied as a value in a context where it does not actually match the required function type, but can be coerced to a function that does. For example a map of type map(xs:integer, xs:string) can be coerced to a function of type function(xs:integer) as xs:string; in this situation a type error will occur only if a call on the function actually returns an empty sequence.

Because of the rules for subtyping of function types according to their signature, it follows that the item type function(A) as item()*, where A is an atomic type, also matches any map, regardless of the type of the keys actually found in the map. For example, a map whose keys are all strings can be supplied where the required type is function(xs:integer) as item()*; a call on the map that treats it as a function with an integer argument will always succeed, and will always return an empty sequence.

The function signature of a map matching type map(K, V), treated as a function, is function(xs:anyAtomicType) as V?. It is thus always a subtype of function(xs:anyAtomicType) as item()* regardless of the actual types of the keys and values in the map. The rules for function coercion mean that any map can be supplied as a value in a context where the required type has a more specific return type, such as function(xs:anyAtomicType) as xs:integer, even when the map does not match in the sense required to satisfy the instance of operator. In such cases, a type error will occur only if an actual call on the map (treated as a function) returns a value that is not an instance of the required return type.

Examples:

  • $M instance of fn(*) returns true()

  • $M instance of fn(xs:anyAtomicType) as item()* returns true()

  • $M instance of fn(xs:integer) as item()* returns true()

  • $M instance of fn(xs:int) as item()* returns true()

  • $M instance of fn(xs:string) as item()* returns true()

  • not($M instance of fn(xs:integer) as xs:string) returns true()

Note:

The last case might seem surprising; however, function coercion ensures that $M can be used successfully anywhere that the required type is fn(xs:integer) as xs:string.

Rules defining whether one map type is a subtype of another are given in 3.3.2.8 Maps.

3.2.8.3 Record Type

Changes in 4.0  

  1. Record types are added as a new kind of ItemType, constraining the value space of maps.

  2. The syntax record(*) is allowed; it matches any map.   [Issue 52 PR 728 10 October 2023]

  3. The syntax record() is allowed; the only thing it matches is an empty map.   [Issue 1491 ]

RecordType ::= AnyRecordType | TypedRecordType
AnyRecordType ::= "record" "(" "*" ")"
TypedRecordType ::= "record" "(" (FieldDeclaration ** ",") ExtensibleFlag? ")"
FieldDeclaration ::= FieldName "?"? ("as" SequenceType)?
FieldName ::= NCName | StringLiteral
ExtensibleFlag ::= "," "*"

A RecordType matches maps that meet specific criteria.

For example, the RecordType record(r as xs:double, i as xs:double) matches a map if the map has exactly two entries: an entry with key "r" whose value is a singleton xs:double value, and an entry with key "i" whose value is also a singleton xs:double value.

If the list of fields ends with ",*" then the record type is said to be extensible. For example, the RecordType record(e as element(Employee), *) matches a map if it has an entry with key "e" whose value matches element(Employee), regardless what other entries the map might contain.

For generality:

  • The syntax record() defines a record type that has no explicit fields and that is not extensible. The only thing it matches is an empty map.

  • The syntax record(*) defines an extensible record type that has no explicit field declarations. It is equivalent to the item type map(*): that is, it allows any map.

A record type can constrain only those entries whose keys are strings, but when the record type is marked as extensible, then other entries may be present in the map with either string or non-string keys. Entries whose key is a string can be expressed using an (unquoted) NCName if the key conforms to NCName syntax, or using a (quoted) string literal otherwise.

Note:

Lookup expressions have been extended in 4.0 so that non-NCName keys can be used without parentheses: employee?"middle name"

If the type declaration for a field is omitted, then item()* is assumed: that is, the map entry may have any type.

If the field name is followed by a question mark, then the value must have the specified type if it is present, but it may also be absent. For example, the RecordType record(first as xs:string, middle? as xs:string, last as xs:string, *) requires the map to have string-valued entries with keys "first" and "last"; it also declares that if the map has an entry with key "middle", the value of that entry must be a single xs:string. Declaring the type as record(first as xs:string, middle? as xs:string?, last as xs:string, *) also allows the entry with key "middle" to be present but empty.

Note:

Within an extensible record type, a FieldDeclaration that is marked optional and has no declared type does not constrain the map in any way, so it serves no practical purpose, but it is permitted because it may have documentary value.

The names of the fields in a record type must be distinct [err:XPST0021].

Record types describe a subset of the value space of maps. They do not define any new kinds of values, or any additional operations. They are useful in many cases to describe more accurately the type of a variable, function parameter, or function result, giving benefits both in the readability of the code, and in the ability of the processor to detect and diagnose type errors and to optimize execution.

If a variable $rec is known to conform to a particular record type, then when a lookup expression $rec?field is used, (a) the processor can report a type error if $rec cannot contain an entry with name field (see 4.13.3.4 Implausible Lookup Expressions), and (b) the processor can make static type inferences about the type of value returned by $rec?field.

Note:

(TODO: change function signatures as suggested here!) A number of functions in the standard function library use maps as function arguments; this is a useful technique where the information to be supplied across the interface is highly variable. However, the type signature for such functions typically declares the argument type as map(*), which gives very little information (and places very few constraints) on the values that are actually passed across. Using record types offers the possibility of improving this: for example, the options argument of fn:parse-json, previously given as map(*), can now be expressed as record(liberal? as xs:boolean, duplicates? as xs:string, escape? as xs:boolean, fallback as fn(xs:string) as xs:string, *). In principle the xs:string type used to describe the duplicates option could also be replaced by a schema-defined subtype of xs:string that enumerates the permitted values ("reject", "use-first", "use-last").

The use of a record type in the signature of such a function causes the coercion rules to be invoked. So, for example, if the function expects an entry in the map to be an xs:double value, it becomes possible to supply a map in which the corresponding entry has type xs:integer.

Greater precision in defining the types of such arguments also enables better type checking, better diagnostics, better optimization, better documentation, and better syntax-directed editing tools.

Note:

One of the motivations for introducing record types is to enable better pattern matching in XSLT when processing JSON input. With XML input, patterns are often based around XML element names. JSON has no direct equivalent of XML’s element names; matching a JSON object such as {longitude: 130.2, latitude: 53.4} relies instead on recognizing the property names appearing in the object. XSLT 4.0, by integrating record types into pattern matching syntax, allows such an object to be matched with a pattern of the form match="record(longitude, latitude)"

Rules defining whether one record type is a subtype of another are given in 3.3.2.10 Record Types.

3.2.8.4 Recursive Record Types

A named record type N is said to be recursive if its definition includes a direct or indirect reference to N.

For example, the following XQuery declaration defines a linked list:

declare record my:list (value as item()*, next? as my:list);

The equivalent in XSLT is:

<xsl:item-type name="my:list" 
               as="record(value as item()*, next? as my:list)"/>

Instances of recursive record types can be constructed and interrogated in the normal way. For example a list of length 3 can be constructed as:

{ "value": 1, "next": { "value": 2, "next": { "value": 3 } } }

and the third value in the map can be retrieved as $list?next?next?value. In practice, recursive data structures are usually manipulated using recursive functions.

It is possible to define a recursive record type that cannot be instantiated, because it has no finite instances: for example (in XQuery) declare record X (next as X);. Such a record declaration is implausible, so the processor may treat it as an error [err:XPST0023], but it is not obliged to do so.

Note:

For an example of a practical use of recursive record types, see the specification of the function fn:random-number-generator.

Recursive type definitions need to be handled specially by the subtyping rules; a naïve approach of simply replacing each reference to a named item type with its definition would make the assessment of the subtype relationship non-terminating. For details see 3.3.2 Subtypes of Item Types.

Example: A Binary Tree

A record used to represent a node in a binary tree might be represented (using XQuery syntax) as:

declare record t:binary-tree 
   ( left? as t:binary-tree, 
     value as item()*, 
     right? as t:binary-tree
   )

A recursive function to walk this tree and enumerate all the values in depth-first order might be written (again using XQuery syntax) as:

declare function t:values($tree as t:binary-tree?) as item()* {
  $tree ! (t:values(?left), ?value, t:values(?right))   
}

 

Example: An Arbitrary Tree

A record used to represent a node in a tree where each node has an arbitrary number of children might be represented (using XQuery syntax) as:

declare record t:tree as (value, children as t:tree*);

A function to walk this tree and enumerate all the values in order might be written as:

declare function t:flatten($tree as t:tree) as item()* {
  $tree?value, $tree?children ! t:flatten(.))   
}

 

Example: Mutually Recursive Types

The usual textbook example of mutually-recursive types is that of a forest consisting of a list of trees, where each tree is a record comprising a value and a forest. As the previous example shows, this structure can be defined straightforwardly in XPath 4.0 without recourse to mutual recursion.

A more realistic example where mutual recursion is needed is for the schema component model used in [XML Schema 1.0] or [XML Schema 1.1]. Simplifying greatly, the data representing an element declaration in XSD may contain references to a complex type, which in turn will typically contain references to further element declarations. The structure therefore involves mutual recursion. A simplified version of the schema component model might be written (in part) as:

declare record ElementDeclaration (
   name as xs:NCName,
   targetNamespace? as xs:anyURI,
   typeDefinition as (SimpleTypeDefinition | ComplexTypeDefinition),
   nillable as xs:boolean,
   abstract as xs:boolean
);
declare record SimpleTypeDefinition (
   name? as xs:NCName,
   targetNamespace? as xs:anyURI,
   baseType? as SimpleTypeDefinition,
   variety as enum("atomic", "list", "union"),
   facets as Facet*,
);
declare record ComplexTypeDefinition (
   name? as xs:NCName,
   targetNamespace? as xs:anyURI,
   baseType? as ComplexTypeDefinition,
   derivationMethod as enum("extension", "restriction"),
   contentType as record (
      variety as enum("empty", "simple", "element-only", "mixed"),
      particle? as Particle,
      simpleTypeDefinition? as SimpleTypeDefinition
   )
);
declare record Particle (
   minOccurs as xs:nonNegativeInteger,
   maxOccurs as (xs:positiveInteger | enum("unbounded")),
   term as (ElementDeclaration | Wildcard | Group)
);
3.2.8.5 Array Type
ArrayType ::= AnyArrayType | TypedArrayType
AnyArrayType ::= "array" "(" "*" ")"
TypedArrayType ::= "array" "(" SequenceType ")"

The AnyArrayType array(*) matches any array. The TypedArrayType array(X) matches any array in which every array member matches the SequenceType X.

Examples:

  • [ 1, 2 ] instance array(*) returns true()

  • [] instance of array(xs:string) returns true()

  • [ "foo" ] instance of array(xs:string) returns true()

  • [ "foo" ] instance of array(xs:integer) returns false()

  • [ (1, 2), (3, 4) ] instance of array(xs:integer) returns false()

  • [ (1, 2), (3, 4) ] instance of array(xs:integer+) returns true()

An array also matches certain other item types, including:

  • item()

  • function(*)

  • function(xs:integer) as item()*

An array that matches array(T) also matches the function type function(xs:integer) as T.

Note:

To understand this rule, consider the use of an array $A in a function call $A($I), which is equivalent to the function call array:get($A, $I). This function accepts any integer for the argument $I, and the result will either be an instance of T, or an error.

The transitivity rules for item type matching mean that if an item A matches a type T, and T is a subtype of U, then A also matches type U. So the fact that an array of strings (array(xs:string)) matches function(xs:integer) as xs:string means that it will also match other function types such as function(xs:long) as item()*.

Furthermore, the rules for function coercion mean that any array can be supplied as a value in a context where it does not actually match the required function type, but can be coerced to a function that does. For example an array of type array(node()) can be coerced to a function of type function(xs:integer) as element(); in this situation a type error will occur only if a call on the function actually returns a node that is not an element node.

The function signature of an array matching array(X), treated as a function, is function(xs:integer) as X. It is thus always a subtype of function(xs:integer) as item()* regardless of the actual member types in the array. The rules for function coercion mean that any array can be supplied as a value in a context where the required type has a more specific return type, such as function(xs:integer) as xs:integer, even when the array does not match in the sense required to satisfy the instance of operator. In such cases, a type error will occur only if an actual call on the array (treated as a function) returns a value that is not an instance of the required return type.

Rules defining whether one array type is a subtype of another are given in 3.3.2.9 Arrays.

3.2.9 xs:error

The type xs:error has an empty value space; it never appears as a dynamic type or as the content type of a dynamic element or attribute type. It was defined in XML Schema in the interests of making the type system complete and closed, and it is also available in XPath 4.0 for similar reasons.

Note:

Even though it cannot occur in an instance, xs:error is a valid type name in a sequence type. The practical uses of xs:error as a sequence type are limited, but they do exist. For instance, an error-handling function that always raises a dynamic error never returns a value, so xs:error is a good choice for the return type of the function.

The semantics of xs:error are well defined as a consequence of the fact that xs:error is defined as a union type with no member types. For example:

  • $x instance of xs:error always returns false, regardless of the value of $x.

  • $x cast as xs:error fails dynamically with error [err:FORG0001]FO40, regardless of the value of $x.

  • $x cast as xs:error? raises a dynamic error [err:FORG0001]FO40 if exists($x) returns true, and evaluates to the empty sequence if empty($x) returns true.

  • xs:error($x) has the same semantics as $x cast as xs:error? (see the previous bullet point)

  • $x castable as xs:error evaluates to false, regardless of the value of $x.

  • $x treat as xs:error raises a dynamic error [err:XPDY0050] if evaluated, regardless of the value of $x. It never fails statically.

All of the above examples assume that $x is actually evaluated. The rules specified in 2.4.4 Errors and Optimization permit an implementation to avoid evaluating $x if the result of the query does not depend upon the value of $x and thus to avoid raising an error.

3.3 Subtype Relationships

Changes in 4.0  

  1. The presentation of the rules for the subtype relationship between sequence types and item types has been substantially rewritten to improve clarity; no change to the semantics is intended.  [Issue 196 PR 202 25 October 2022]

[Definition: Given two sequence types or item types, the rules in this section determine if one is a subtype of the other. If a type A is a subtype of type B, it follows that every value matched by A is also matched by B.]

Note:

The relationship subtype(A, A) is always true: every type is a subtype of itself.

Note:

The converse is not necessarily true: we cannot infer that if every value matched by A is also matched by B, then A is a subtype of type B. For example, A might be defined as the set of strings matching the regular expression [A-Z]*, while B is the set of strings matching the regular expression [A-Za-z]*; no subtype relationship holds between these types.

The rules for deciding whether one sequence type is a subtype of another are given in 3.3.1 Subtypes of Sequence Types. The rules for deciding whether one item type is a subtype of another are given in 3.3.2 Subtypes of Item Types.

Note:

The subtype relationship is not acyclic. There are cases where subtype(A, B) and subtype(B, A) are both true. This implies that A and B have the same value space, but they can still be different types. For example this applies when A is a union type with member types xs:string and xs:integer, while B is a union type with member types xs:integer and xs:string. These are different types ("23" cast as A produces a string, while "23" cast as B produces an integer, because casting is attempted to each member type in order) but both types have the same value space.

3.3.1 Subtypes of Sequence Types

We use the notation A ⊑ B, or subtype(A, B) to indicate that a sequence type A is a subtype of a sequence type B. This section defines the rules for deciding whether any two sequence types have this relationship.

To define the rules, we divide sequence types into six categories:

  • The category empty includes the sequence types empty-sequence(), xs:error* and xs:error?. All these sequence types match the empty sequence as their only instance.

  • The category void includes the sequence types xs:error and xs:error+, which have no instances (not even the empty sequence).

  • The categories X?, X*, X and X+ includes all sequence types having an item type X other than xs:error, together with an occurrence indicator of ? (zero or more), * (one or more), absent (exactly one), or + (one or more) respectively. We use the notation Xi to indicate the item type of such a sequence type.

The judgement A ⊑ B is then determined by the categories of the two sequence types, as defined in the table below. In many cases this depends on the relationship between the item types of A and B. This is denoted using the notation AiBi , as defined in 3.3.2 Subtypes of Item Types.

Sequence type B
empty Bi? Bi* Bi Bi+ void
Sequence type A empty true true true false false false
Ai? false AiBi AiBi false false false
Ai* false false AiBi false false false
Ai false AiBi AiBi AiBi AiBi false
Ai+ false false AiBi false AiBi false
void true true true true true true

3.3.2 Subtypes of Item Types

We use the notation A ⊆ B, or itemtype-subtype(A, B) to indicate that an item type A is a subtype of an item type B. This section defines the rules for deciding whether any two item types have this relationship.

The rules in this section apply to item types, not to item type designators. For example, if the name STR has been defined in the static context as a named item type referring to the type xs:string, then anything said here about the type xs:string applies equally whether it is designated as xs:string or as STR, or indeed as the parenthesized forms (xs:string) or (STR).

References to named item types are handled as described in 3.3.2.11 Named Item Types.

The relationship A ⊆ B is true if and only if at least one of the conditions listed in the following subsections applies:

3.3.2.1 General Rules

Given item types A and B, AB is true if any of the following apply:

  1. A is xs:error.

  2. B is item().

  3. A and B are the same item type.

  4. There is an item type X such that AX and XB . (This is referred to below as the transitivity rule).

Note:

The first rule is technically redundant: it is implied by the second rule in 3.3.2.3 Atomic and Union Types. The type xs:error is defined as a union type with no member types; therefore it is automatically true that every member type T satisfies TB .

3.3.2.2 Choice Item Types

The following rules determine whether AB is true in the case where either A or B or both is a choice item type.

Firstly, if one of the operands is not a choice item type, then it is treated as a choice item type with a single member type. The rule is then:

AB is true if for every member type a in A, there is a member type b in B such that ab .

For example, (xs:int | xs:long) is a subtype of (xs:decimal | xs:date) because both xs:int and xs:long are subtypes of xs:decimal.

Note:

Because an enumeration type is defined as a choice type of singleton enumerations, these rules have the consequence, for example, that enum("A", "B") is a subtype of enum("A", "B", "C").

Note:

The type xs:int is not a subtype of (xs:negativeInteger | xs:nonNegativeInteger), because it does not satisfy this rule. This is despite the fact that the value space of xs:int is a subset of the value space of (xs:negativeInteger | xs:nonNegativeInteger).

3.3.2.3 Atomic and Union Types

Given item types A and B, AB is true if any of the following apply:

  1. A and B are generalized atomic types, and derives-from(A, B) returns true.

    The derives-from relationship is defined in 3.1.2 Sequence Type Matching.

    Examples:
    • xs:integer ⊆ xs:decimal because xs:integer is derived by restriction from xs:decimal.

    • xs:decimal ⊆ xs:numeric because xs:numeric is a pure union type that includes xs:decimal as a member type.

    • enum("red") ⊆ xs:string because the singleton enumeration type enum("red") is defined to be an atomic type derived from xs:string.

    • enum("red") ⊆ enum("red", "green") because the enumeration type enum("red", "green") is defined to be a union type that has the generalized atomic type enum("red") as a member type.

  2. A is a pure union type, and every type T in the transitive membership of A satisfies TB .

    Examples:
    • (xs:short | xs:long) ⊆ xs:integer because xs:short ⊆ xs:integer and xs:long ⊆ xs:integer.

    • (P | Q) ⊆ (P | Q | R) because P ⊆ (P | Q | R) and Q ⊆ (P | Q | R).

    • enum("red", "green") ⊆ xs:string because the enumeration type enum("red") ⊆ xs:string and enum("green") ⊆ xs:string.

    • enum("red", "green") ⊆ enum("red", "green", "blue") because enum("red") ⊆ enum("red", "green", "blue") and enum("green") ⊆ enum("red", "green", "blue").

    • enum("red", "green", "blue") ⊆ (enum("red", "green") | enum("blue")) because each of the types enum("red"), enum("green"), and enum("blue") is a subtype of one of the two members of the union type.

3.3.2.4 Node Types: General Rules

Given item types A and B, AB is true if any of the following apply:

  1. A is a KindTest and B is node().

    Example:

    comment() ⊆ node()

  2. A is processing-instruction(N) for any name N, and B is processing-instruction().

    Example:

    processing-instruction('pi') ⊆ processing-instruction()

  3. A is document-node(E) for any E, and B is document-node().

    Examples:

    document-node(element(chap)) ⊆ document-node()

    document-node(*) ⊆ document-node()

  4. All the following are true, after expanding document-node(NameTestUnion) to document-node(element(NameTestUnion)):

    1. A is document-node(Ae)

    2. B is document-node(Be)

    3. AeBe

    Examples:

    document-node(element(title)) ⊆ document-node(element(*)).

    document-node(title) ⊆ document-node(*).

3.3.2.5 Node Types: Element Tests

[Definition: In these rules, if MU and NU are NameTestUnions, then MU wildcard-matches NU is true if every name that matches MU also matches NU.]

More specifically, this is the case if for every NameTest M in MU there is a NameTest N in NU where at least one of the following applies:

  1. M and N are the same NameTest.

  2. M is an EQName and N is a Wildcard that matches M.

  3. N is the Wildcard *.

Given item types A and B, AB is true if any of the following apply.

  1. A is an ElementTest and B is either element() or element(*)

  2. All the following are true:

    1. A is either element(An) or element(An, T) or element(An, T?) for any type T

    2. B is either element(Bn) or element(Bn, xs:anyType?)

    3. An wildcard-matches Bn

    Examples:
    • element(title) ⊆ element(*)

    • element(title, xs:string) ⊆ element(*)

    • element(title|heading, xs:string) ⊆ element(*)

    • element(title, xs:string) ⊆ element(title|heading)

    • element(title, xs:string?) ⊆ element(*)

    • element(title|heading, xs:string) ⊆ element(*)

    • element(title) ⊆ element(title, xs:anyType?)

    • element(title, xs:integer) ⊆ element(title|heading, xs:anyType?)

    • element(title, xs:string?) ⊆ element(title, xs:anyType?)

    • element(my:title|your:title) ⊆ element(*:title)

    • element(my:title|my:heading) ⊆ element(my:*)

  3. All the following are true:

    1. A is element(An, At)

    2. B is element(Bn, Bt)

    3. An wildcard-matches Bn

    4. derives-from(At, Bt).

    Examples:
    • element(size, xs:integer) ⊆ element(size, xs:decimal)

    • element(size, xs:integer) ⊆ element(size|größe, xs:decimal)

    • element(size, xs:integer) ⊆ element(*, xs:decimal)

    • element(*, xs:integer) ⊆ element(*, xs:decimal)

    • element(my:*, xs:integer) ⊆ element(*, xs:decimal)

  4. All the following are true:

    1. A is either element(An, At) or element(An, At?)

    2. B is element(Bn, Bt?)

    3. An wildcard-matches Bn

    4. derives-from(At, Bt).

    Examples:
    • element(size, xs:integer) ⊆ element(size, xs:decimal?)

    • element(size, xs:integer?) ⊆ element(*, xs:decimal?)

    • element(*, xs:integer) ⊆ element(*, xs:decimal?)

    • element(my:*, xs:integer?) ⊆ element(*, xs:decimal?)

  5. All the following are true:

    1. A is schema-element(An)

    2. B is schema-element(Bn)

    3. Every element declaration that is an actual member of the substitution group of An is also an actual member of the substitution group of Bn.

    Note:

    The fact that P is a member of the substitution group of Q does not mean that every element declaration in the substitution group of P is also in the substitution group of Q. For example, Q might block substitution of elements whose type is derived by extension, while P does not.

  6. A is element(A1|A2|..., T) (where T may be absent), and for each An, element(An, T) ⊆ B .

    Examples:
    • element(a|b) ⊆ (element(a)|element(b)|element(c))

    • element(a|b, xs:integer) ⊆ (element(a, xs:decimal) | element(b, xs:integer))

3.3.2.6 Node Types: Attribute Tests

Given item types A and B, AB is true if any of the following apply:

  1. A is an AttributeTest and B is either attribute() or attribute(*)

  2. All the following are true:

    1. A is either attribute(An) or attribute(An, T) for any type T.

    2. B is either attribute(Bn) or attribute(Bn, xs:anyAtomicType)

    3. An wildcard-matches Bn

    Examples:
    • attribute(code) ⊆ attribute(*)

    • attribute(code|status) ⊆ attribute(*)

    • attribute(code, xs:untypedAtomic) ⊆ attribute(*)

    • attribute(code|status, xs:string) ⊆ attribute(code, xs:anyAtomicType)

    • attribute(my:code|your:code) ⊆ attribute(*:code)

    • attribute(my:code|my:status) ⊆ attribute(my:*)

  3. All the following are true:

    1. A is attribute(An, At)

    2. B is attribute(Bn, Bt)

    3. An wildcard-matches Bn

    4. derives-from(At, Bt).

    Examples:
    • attribute(*, xs:ID) ⊆ attribute(*, xs:string)

    • attribute(my:*, xs:ID) ⊆ attribute(*, xs:string)

    • attribute(code, xs:ID) ⊆ attribute(code|status, xs:string)

    • attribute(code, xs:ID) ⊆ attribute(*, xs:string)

    • attribute(code, xs:ID) ⊆ attribute(*:code, xs:ID)

    • attribute(my:code|my:status, xs:ID) ⊆ attribute(my:*, xs:string)

  4. All the following are true:

    1. A is schema-attribute(An)

    2. B is schema-attribute(Bn)

    3. the expanded QName of An equals the expanded QName of Bn

  5. A is attribute(A1|A2|..., T) (where T may be absent), and for each An, attribute(An, T) ⊆ B .

    Examples:
    • attribute(a|b) ⊆ (attribute(a)|attribute(b)|attribute(c))

    • attribute(a|b, xs:integer) ⊆ (attribute(a, xs:decimal) | attribute(b))

3.3.2.7 Functions

Given item types A and B, AB is true if any of the following apply:

  1. All the following are true:

    1. A is a FunctionType

    2. B is function(*)

    Example:

    function(xs:integer) as xs:string ⊆ function(*)

  2. All the following are true:

    1. A is function(a1, a2, ... aM) as RA

    2. B is function(b1, b2, ... bN) as RB

    3. N (the arity of B) equals M (the arity of A)

    4. RARB

    5. For all values of p between 1 and N, bpap

    Examples:
    • function(xs:integer) as xs:string ⊆ function(xs:long) as xs:string

    • function(xs:integer) as xs:ID ⊆ function(xs:integer) as xs:string

    • function(xs:integer) as xs:ID ⊆ function(xs:long) as xs:string

    Note:

    Function return types are covariant because this rule requires RARB for return types. Function parameter types are contravariant because this rule requires bpap for parameter types.

3.3.2.8 Maps

Given item types A and B, AB is true if any of the following apply:

  1. Both of the following are true:

    1. A is map(K, V), for any K and V

    2. B is map(*)

    Example:

    map(xs:integer, item()*) ⊆ map(*)

  2. All the following are true:

    1. A is map(Ka, Va)

    2. B is map(Kb, Vb)

    3. KaKb

    4. VaVb

    Example:

    map(xs:long, item()) ⊆ map(xs:integer, item()+)

  3. Both the following are true:

    1. A is map(*) (or, because of the transitivity rules, any other map type)

    2. B is function(*)

    Example:

    map(xs:long, xs:string?) ⊆ function(*)

  4. Both the following are true:

    1. A is map(*) (or, because of the transitivity rules, any other map type)

    2. B is function(xs:anyAtomicType) as item()*

    Example:

    map(xs:long, xs:string?) ⊆ function(xs:anyAtomicType) as item()*

  5. All the following are true:

    1. A is map(K, V)

    2. B is function(xs:anyAtomicType) as R

    3. VR

    4. empty-sequence()R

    Examples:
    • map(xs:int, node()) ⊆ function(xs:anyAtomicType) as node()?

    • map(xs:int, node()+) ⊆ function(xs:anyAtomicType) as node()*

    The function accepts type xs:anyAtomicType rather than xs:int, because $M("xyz") is a valid call on a map (treated as a function) even when all the keys in the map are integers.

    The return type of the function is extended from node() or node()+ to allow an empty sequence because $M("xyz") can return an empty sequence even if none of the entries in the map contains an empty sequence.

3.3.2.9 Arrays

Given item types A and B, AB is true if any of the following apply:

  1. Both the following are true:

    1. A is array(X)

    2. B is array(*)

    Example:

    array(xs:integer) ⊆ array(*)

  2. All the following are true:

    1. A is array(X)

    2. B is array(Y)

    3. XY

    Example:

    array(xs:integer) ⊆ array(xs:decimal+)

  3. Both the following are true:

    1. A is array(*) (or, because of the transitivity rules, any other array type)

    2. B is function(*)

    Example:

    array(xs:integer) ⊆ function(*)

  4. Both the following are true:

    1. A is array(*) (or, because of the transitivity rules, any other array type)

    2. B is function(xs:integer) as item()*

    Example:

    array(*) ⊆ function(xs:integer) as item()*

  5. Both the following are true:

    1. A is array(X)

    2. B is function(xs:integer) as X

    Example:

    array(xs:string) ⊆ function(xs:integer) as xs:string

3.3.2.10 Record Types

Given item types A and B, A B is true if any of the following apply:

  1. A is map(*) and B is record(*).

  2. All of the following are true:

    1. A is a record type.

    2. B is map(*) or record(*).

    Examples:

    record(longitude, latitude)map(*)

    record(longitude, latitude, *)record(*)

    record(*)map(*)

  3. All of the following are true:

    1. A is a non-extensible record type

    2. B is map(K, V)

    3. K is either xs:string or xs:anyAtomicType

    4. For every field F in A, where T is the declared type of F (or its default, item()*), TV .

    Examples:
    • record(x, y)map(xs:string, item()*)

    • record(x as xs:double, y as xs:double)map(xs:string, xs:double)

  4. All of the following are true:

    1. A is a non-extensible record type.

    2. B is a non-extensible record type.

    3. Every field in A is also declared in B.

    4. Every mandatory field in B is also declared as mandatory in A.

    5. For every field that is declared in both A and B, where the declared type in A is T and the declared type in B is U, TU .

    Examples:
    • record(x, y as xs:integer) ⊆ record(x, y as xs:decimal)

    • record(x, y) ⊆ record(x, y, z?)

  5. All of the following are true:

    1. A is an extensible record type

    2. B is an extensible record type

    3. Every mandatory field in B is also declared as mandatory in A.

    4. For every field that is declared in both A and B, where the declared type in A is T and the declared type in B is U, TU .

    5. For every field that is declared in B but not in A, the declared type in B is item()*.

    Examples:
    • record(x, y, z, *) ⊆ record(x, y, *)

    • record(x?, y?, z?, *) ⊆ record(x, y, *)

    • record(x as xs:integer, y as xs:integer, *) ⊆ record(x as xs:decimal, y as xs:integer*, *)

    • record(x as xs:integer, *) ⊆ record(x as xs:decimal, y as item(), *)

  6. All of the following are true:

    1. A is a non-extensible record type.

    2. B is an extensible record type.

    3. Every mandatory field in B is also declared as mandatory in A.

    4. For every field that is declared in both A and B, where the declared type in A is T and the declared type in B is U, TU .

    Examples:
    • record(x, y as xs:integer) ⊆ record(x, y as xs:decimal, *)

    • record(y as xs:integer) ⊆ record(x?, y as xs:decimal, *)

3.3.2.11 Named Item Types

This section describes how references to named item types are handled when evaluating the subtype relationship.

Named item types can be classified as recursive or non-recursive. A recursive type is one that references itself, directly or indirectly. Only named record types are allowed to be recursive.

In the case of references to non-recursive named item types, the reference is fully expanded as the first step in evaluating the subtype relationship. For example this means that if U is a named item type with the expansion (xs:integer | xs:double), then xs:integer ⊆ U is true, because xs:integer ⊆ (xs:integer | xs:double) is true.

Recursive record types are considered to be, in the terminology of the computer science literature, iso-recursive (rather than equi-recursive). This means that a recursive type name is not treated as being equivalent to its expansion (at any depth). For example, if the named item type T has the expansion record(A as item()*, B as T?), then the type array(T) is not considered to be equivalent to array(record(A as item()*, B as T?)), despite the fact that the two types have exactly the same instances.

The rules are therefore defined as follows:

  • If B is a reference to a recursive record type, then AB is true if and only if A and B are references to the same named record type.

  • If A is a reference to a recursive named item type, then AB is true if either:

    • A and B are references to the same named record type.

    • record(*) ⊆ B.

      Note:

      This is because only record types are allowed to be recursive.

Note:

The decision to make recursive types iso-recursive rather than equi-recursive was made largely because it saves a great deal of implementation complexity without any serious adverse effects for users. In practice, problems can be avoided by using named record types consistently (for example, avoiding having two named record types with different names but identical definitions).

3.4 Coercion Rules

Changes in 4.0  

  1. The term "function conversion rules" used in 3.1 has been replaced by the term "coercion rules".   [ PR 254 29 November 2022]

  2. The coercion rules allow “relabeling” of a supplied atomic item where the required type is a derived atomic type: for example, it is now permitted to supply the value 3 when calling a function that expects an instance of xs:positiveInteger.   [Issue 117 PR 254 29 November 2022]

  3. The coercion rules now allow any numeric type to be implicitly converted to any other, for example an xs:double is accepted where the required type is xs:double.   [Issue 980 PR 911 30 January 2024]

  4. The coercion rules now allow conversion in either direction between xs:hexBinary and xs:base64Binary.   [Issues 130 480 PR 815 7 November 2023]

[Definition: The coercion rules are rules used to convert a supplied value to a required type, for example when converting an argument of a function call to the declared type of the function parameter. ] The required type is expressed as a sequence type. The effect of the coercion rules may be to accept the value as supplied, to convert it to a value that matches the required type, or to reject it with a type error.

This section defines how the coercion rules operate; the situations in which the rules apply are defined elsewhere, by reference to this section.

Note:

In previous versions of this specification, the coercion rules were referred to as the function conversion rules. The terminology has changed because the rules are not exclusively associated with functions or function calling.

If the required type is empty-sequence(), no coercion takes place (the supplied value must be an empty sequence, or a type error occurs).

In all other cases, the required sequence type T comprises a required item type R and an optional occurrence indicator. The coercion rules are then applied to a supplied value V and the required type T as follows:

  1. If XPath 1.0 compatibility mode is true and V is not an instance of the required type T, then the conversions defined in 3.4.1 XPath 1.0 Compatibility Rules are applied to V. Then:

  2. Each item in V is processed against the required item type R using the item coercion rules defined in 3.4.2 Item Coercion Rules, and the results are sequence-concatenated into a single sequence V′.

  3. A type error is raised if the cardinality of V′ does not match the required cardinality of T [err:XPTY0004].

3.4.1 XPath 1.0 Compatibility Rules

These rules are used to process a value V against a required sequence type T when XPath 1.0 compatibility mode is true.

  1. If the occurrence indicator of T is either absent or ? (examples: xs:string, xs:string?, xs:untypedAtomic, xs:untypedAtomic?, node(), node()?, item(), item()?), then V is effectively replaced by V[1].

  2. If T is xs:string or xs:string?, then V is effectively replaced by fn:string(V).

    Note:

    This rule does not apply where T is derived from xs:string or xs:string?, because derived types did not arise in XPath 1.0.

  3. If T is xs:double or xs:double?, then V is effectively replaced by fn:number(V).

    Note:

    This rule does not apply where T is derived from xs:double or xs:double?, because derived types did not arise in XPath 1.0.

Note:

The special rules for XPath 1.0 compatibility mode are used for converting the arguments of a static function call, and in certain XSLT constructs. They are not invoked in other contexts such as dynamic function calls, for converting the result of an inline function to its required type, for partial function application, or for implicit function calls such as occur when evaluating functions such as fn:for-each and fn:filter.

3.4.2 Item Coercion Rules

The rules in this section are used to process each item J in a supplied sequence, given a required item type R.

  1. If R is a generalized atomic type (for example, if it is an atomic type, a pure union type, or an enumeration type), and J is not an atomic item, then:

    1. J is atomized to produce a sequence of atomic items JJ.

    2. Each atomic item in JJ is coerced to the required type R by recursive application of the item coercion rules (the rules in this section) to produce a value V.

    3. The result is the sequence-concatenation of the V values.

    Note:

    For example, if J is an element with type annotation xs:integer, and R is the union type xs:numeric, then the effect is to atomize the element to an xs:integer, and then to coerce the resulting xs:integer to xs:numeric (which leaves the integer unchanged). This is not the same as attempting to coerce the element to each of the alternatives of the union type in turn, which would deliver an instance of xs:double.

  2. Otherwise, if R is a choice item type or a pure union type (which includes the case where it is an enumeration type), then:

    1. If J matches (is an instance of) one of the alternatives in R, then:

      1. If the first alternative in R that J matches is a typed function type (see 3.2.8.1 Function Type), then function coercion is applied to coerce J to that function type, as described in 3.4.4 Function Coercion.

      2. Otherwise, J is used as is.

    2. Otherwise, the item coercion rules (the rules in this section) are applied to J recursively with R set to each of the alternatives in the choice or union item type, in order, until an alternative is found that does not result in a type error; a type error is raised only if all alternatives fail.

      The error code used in the event of failure should be the error code arising from the first unsuccessful matching attempt. (The diagnostic information associated with the error may also describe how further attempts failed.)

    Note:

    Suppose the required type is (xs:integer | element(e))* and the supplied value is the sequence (<e>22</e>, 23, <f>24</f>). Item coercion is applied independently to each of the three items in this sequence. The first item matches one of the alternatives, namely element(e), so it is returned unchanged as an element node. The second item (the integer 23) also matches one of the alternatives, and is returned unchanged as an integer. The third item does not match any of the alternatives, so coercion is attempted to each one in turn. Coercion to type xs:integer succeeds (by virtue of atomization and untyped atomic conversion), so the final result is the sequence (<e>22</e>, 23, 24)

    Note:

    Suppose the required type is enum("red", "green", "blue") and the supplied value is "green". The enumeration type is defined as a choice item type whose alternatives are singleton enumerations, so the rules are applied first to the type enum("red") (which fails), and then to the type enum("green") (which succeeds). The strings in an enumeration type are required to be distinct so the order of checking is in this case immaterial. The supplied value will be accepted, and will be relabeled as an instance of enum("green"), which is treated as a schema type equivalent to a type derived from xs:string by restriction.

    Note:

    Schema-defined union types behave in exactly the same way as choice item types.

  3. If R is an atomic type and J is an atomic item, then:

    1. If J is an instance of R then it is used unchanged.

    2. If J is an instance of type xs:untypedAtomic then:

      1. If R is an enumeration type then A is cast to xs:string.

      2. If R is namespace-sensitive then a type error [err:XPTY0117] is raised.

    3. Otherwise, J is cast to type R.

  4. If there is an entry (from, to) in the following table such that J is an instance of from, and to is R, then J is cast to type R.

    Implicit Casting
    from to
    xs:decimal xs:double
    xs:double xs:decimal
    xs:decimal xs:float
    xs:float xs:decimal
    xs:float xs:double
    xs:double xs:float
    xs:string xs:anyURI
    xs:anyURI xs:string
    xs:hexBinary xs:base64Binary
    xs:base64Binary xs:hexBinary

    Note:

    The item type in the to column must match R exactly; however, J may belong to a subtype of the type in the from column.

    For example, an xs:NCName will be cast to type xs:anyURI, but an xs:anyURI will not be cast to type xs:NCName.

    Similarly, an xs:integer will be cast to type xs:double, but an xs:double will not be cast to type xs:integer.

  5. If R is derived from some primitive atomic type P, then J is relabeled as an instance of R if it satisfies all the following conditions:

    • J is an instance of P.

    • J is not an instance of R.

    • The datumDM40 of J is within the value space of R.

    Relabeling an atomic item changes the type annotation but not the datumDM40. For example, the xs:integer value 3 can be relabeled as an instance of xs:unsignedByte, because the datum is within the value space of xs:unsignedByte.

    Note:

    Relabeling is not the same as casting. For example, the xs:decimal value 10.1 can be cast to xs:integer, but it cannot be relabeled as xs:integer, because its datum not within the value space of xs:integer.

    Note:

    The effect of this rule is that if, for example, a function parameter is declared with an expected type of xs:positiveInteger, then a call that supplies the literal value 3 will succeed, whereas a call that supplies -3 will fail.

    This differs from previous versions of this specification, where both these calls would fail.

    This change allows the arguments of existing functions to be defined with a more precise type. For example, the $position argument of array:get could be defined as xs:positiveInteger rather than xs:integer.

    Note:

    If T is a union type with members xs:negativeInteger and xs:positiveInteger)* and the supplied value is the sequence (20, -20), then the effect of these rules is that the first item 20 is relabeled as type xs:positiveInteger and the second item -20is relabeled as type xs:negativeInteger.

    Note:

    Promotion (for example of xs:float to xs:double) occurs only when T is a primitive type. Relabeling occurs only when T is a derived type. Promotion and relabeling are therefore never combined.

    Note:

    A singleton enumeration type such as enum("green") is treated as an atomic type derived by restriction from xs:string; so if the xs:string value "green" is supplied in a context where the required type is enum("red", "green", "blue"), the value will be accepted and will be relabeled as an instance of enum("green").

  6. If R is an ArrayType other than array(*) and J is an array, then J is converted to a new array by converting each member to the required member type by applying the coercion rules recursively.

    Note:

    For example, if the required type is array(xs:double) and the supplied value is [ 1, 2 ], the array is converted to [ 1e0, 2e0 ].

  7. If R is a MapType other than map(*) and J is a map, then J is converted to a new map as follows:

    1. Each key in the supplied map is converted to the required map key type by applying the coercion rules. If the resulting map would contain duplicate keys, a type error is raised [err:XPTY0004].

    2. The corresponding value is converted to the required map value type by applying the coercion rules recursively.

    Note:

    For example, if the required type is map(xs:string, xs:double) and the supplied value is { "x": 1, "y": 2 }, the map is converted to { "x": 1e0, "y": 2e0 }.

    Note:

    Duplicate keys can occur if the value space of the target type is more restrictive than the original type. For example, an error is raised if the map { 1.2: 0, 1.2000001: 0 }, which contains two keys of type xs:decimal, is coerced to the type map(xs:float, xs:integer).

  8. If R is a RecordType and J is a map, then J is converted to a new map as follows:

    1. The keys in the supplied map are unchanged.

    2. In any map entry whose key is an xs:string equal to the name of one of the field declarations in R, the corresponding value is converted to the required type defined by that field declaration, by applying the coercion rules recursively (but with XPath 1.0 compatibility mode treated as false).

    Note:

    For example, if the required type is record(longitude as xs:double, latitude as xs:double) and the supplied value is { "longitude": 0, "latitude": 53.2 }, then the map is converted to { "longitude": 0.0e0, "latitude": 53.2e0 }.

  9. If R is a TypedFunctionType and J is a function item, then function coercion is applied to J.

    Note:

    Function coercion applies even if J is already an instance of R.

    Maps and arrays are functions, so function coercion applies to them as well.

  10. If, after the above conversions, the resulting item does not match the expected item type R according to the rules for SequenceType Matching, a type error is raised [err:XPTY0004].

    Note:

    Under the general rules for type errors (see 2.4.1 Kinds of Errors), a processor may report a type error during static analysis if it will necessarily occur when the expression is evaluated. For example, the function call fn:abs("beer") will necessarily fail when evaluated, because the function requires a numeric value as its argument; this may be detected and reported as a static error.

3.4.3 Implausible Coercions

An expression is deemed to be implausible [err:XPTY0006] if the static type of the expression, after applying all necessary coercions, is substantively disjoint with the required type T.

[Definition: Two sequence types are deemed to be substantively disjoint if (a) neither is a subtype of the other (see 3.3.1 Subtypes of Sequence Types) and (b) the only values that are instances of both types are one or more of the following:

  • The empty sequence, ().

  • The empty map, {}.

  • The empty array, [].

]

Note:

Examples of pairs of sequence types that are substantively disjoint include:

  • xs:integer* and xs:string*

  • map(xs:integer, node()) and map(xs:string, node())

  • array(xs:integer) and array(xs:string)

For example, supplying a value whose static type is xs:integer* when the required type is xs:string* is implausible, because it can succeed only in the special case where the actual value supplied is an empty sequence.

Note:

The case where the supplied type and the required type are completely disjoint (for example map(*) and array(*)) is covered by the general rules for type errors: that case can always be reported as a static error.

Examples of implausible coercions include the following:

  • round(timezone-from-time($now)). The result of fn:timezone-from-time is of type xs:dayTimeDuration?, which is substantively disjoint with the required type of fn:round, namely xs:numeric?.

  • function($x as xs:integer) as array(xs:string) { array { 1 to $x } }. The type of the function body is array(xs:integer), which is substantively disjoint with the required type array(xs:string): the function can succeed only in the exceptional case where the function body delivers an empty array.

3.4.4 Function Coercion

Changes in 4.0  

  1. Function coercion now allows a function with arity N to be supplied where a function of arity greater than N is expected. For example this allows the function true#0 to be supplied where a predicate function is required.

  2. It has been clarified that function coercion applies even when the supplied function item matches the required function type. This is to ensure that arguments supplied when calling the function are checked against the signature of the required function type, which might be stricter than the signature of the supplied function item.   [Issue 1020 PRs 1023 1128 9 April 2024]

Function coercion is a transformation applied to function items during application of the coercion rules. [Definition: Function coercion wraps a function item in a new function whose signature is the same as the expected type. This effectively delays the checking of the argument and return types until the function is called.]

Given a function F, and an expected function type T, function coercion proceeds as follows:

  1. If F has higher arity than T, a type error is raised [err:XPTY0004]

  2. If F has lower arity than T, then F is wrapped in a new function that declares and ignores the additional argument; the following steps are then applied to this new function.

    For example, if T is function(node(), xs:boolean) as xs:string, and the supplied function is fn:name#1, then the supplied function is effectively replaced by function($n as node(), $b as xs:boolean) as xs:string { fn:name($n) }

    Note:

    This mechanism makes it easier to design versatile and extensible higher-order functions. For example, in previous versions of this specification, the second argument of the fn:filter function expected an argument of type function(item()) as xs:boolean. This has now been extended to function(item(), xs:integer) as xs:boolean, but existing code continues to work, because callback functions that are not interested in the value of the second argument simply ignore it.

  3. A type error is raised [err:XPTY0004] if, for any parameter type, or for the result type, the relevant type in the signature of the supplied function and the relevant type in the expected function type are substantively disjoint.

    For example, the types xs:integer and xs:string are substantively disjoint, so a function with signature function(xs:integer) as xs:boolean cannot be supplied where the expected type is function(xs:string) as xs:boolean.

  4. Function coercion then returns a new function item with the following properties (as defined in Section 2.9.4 Function ItemsDM40):

    • name: The name of F (if not absent).

    • identity: A new function identity distinct from the identity of any other function item.

      Note:

      See also 4.5.2.7 Function Identity.

    • parameter names: The parameter names of F.

    • signature: Annotations is set to the annotations of F. TypedFunctionType is set to the expected type.

    • implementation: In effect, a FunctionBody that calls F, passing it the parameters of this new function, in order.

    • nonlocal variable bindings: An empty mapping.

These rules have the following consequences:

  • SequenceType matching of the function’s arguments and result are delayed until that function is called.

  • When the coerced function is called, the supplied arguments must match the parameter typed defined in T; it is not sufficient to match the parameter types defined in F.

  • The coercion rules rules applied to the function’s arguments and result are defined by the SequenceType it has most recently been coerced to. Additional coercion rules could apply when the wrapped function is called.

  • If an implementation has static type information about a function, that can be used to type check the function’s argument and return types during static analysis.

  • When function coercion is applied to a map or an array, the resulting function is not a map or array, and cannot be used as such. For example, the expression

    let $f as function(xs:integer) as xs:boolean := { 0: false(), 1: true() }
    return $f?0

    raises a type error, because a lookup expression requires the left hand operand to be a map or array, and $f is neither.

  • When function types are used as alternatives in a choice item type, the supplied function is coerced to the first alternative for which coercion does not raise a type error. In this situation it is important to write the alternatives in order, with the most specific first.

Note:

The semantics of function coercion are specified in terms of wrapping the functions. Static typing may be able to reduce the number of places where this is actually necessary. However, it cannot be assumed that because a supplied function is an instance of the required function type, no function coercion is necessary: the supplied function might not perform all required checks on the types of its arguments.

Since maps and arrays are also functions in XPath 4.0, function coercion applies to them as well. For instance, consider the following expression:

let $m := {
  "Monday" : true(),
  "Wednesday" : false(),
  "Friday" : true()
}
let $days := ("Monday", "Tuesday", "Wednesday", "Thursday", "Friday")
return filter($days, $m)

The map $m is an instance of function(xs:anyAtomicType?) as item()*. When the fn:filter() function is called, the following occurs to the map:

  1. The map $m is treated as a function equivalent to map:get($m, ?).

  2. The coercion rules result in applying function coercion to this function, wrapping it in a new function (M′) with the signature function(item(), xs:integer) as xs:boolean.

  3. When M′ is called by fn:filter(), coercion and SequenceType matching rules are applied to the argument, resulting in an item() value ($a) or a type error.

  4. The function map:get($m, ?) is called with $a as the argument; this returns either an xs:boolean or the empty sequence (call the result R).

  5. The wrapper function $p applies the coercion rules to R. If R is an xs:boolean the matching succeeds. When it is an empty sequence (in particular, $m does not contain a key for "Tuesday"), a type error is raised [err:XPTY0004], since the expected type is xs:boolean and the actual type is an empty sequence.

Consider the following expression:


let $m := {
   "Monday" : true(),
   "Wednesday" : false(),
   "Friday" : true(),
}
let $days := ("Monday", "Wednesday", "Friday")
return filter($days, $m)

In this case the result of the expression is the sequence ("Monday", "Friday"). But if the input sequence included the string "Tuesday", the filter operation would fail with a type error.

Note:

Function coercion applies even if the supplied function matches the required type.

For example, consider this case:

declare function local:filter(
  $s as item()*, 
  $p as function(xs:string) as xs:boolean
) as item()* {
  $s[$p(.)]
};

let $f := function($a) { $a mod 2 = 0 }
return local:filter(1 to 10, $f)

Here the supplied function $f is an instance of the required type, because its signature defaults the argument type to item()*, which is a supertype of xs:string. The expression $s[$p(.)] could in principle succeed. However, function coercion ensures that the supplied function is wrapped in a function that requires the argument to be of type xs:string, so the call fails with a type error when the wrapping function is invoked supplying an xs:integer as the argument.

3.4.5 Examples of Coercions

This section illustrates the effect of the coercion rules with examples.

Example: Coercion to xs:string

Consider the case where the required type (of a variable, or a function argument) is xs:string. For example, the second argument of fn:matches, which expects a regular expression. The table below illustrates the values that might be supplied, and the coercions that are applied.

Supplied Value Coercion
"[0-9]"

None; the supplied value is an instance of the required type.

default-language()

None; the supplied value is an instance of xs:language, which is a subtype of the required type xs:string.

<a>[0-9]</a>

The supplied element node is atomized. Unless it has been schema-validated, the typed value will be an instance of xs:untypedAtomic, which is accepted when the required type is xs:string.

Supplying an element whose type annotation is (say) xs:date will fail with a type error.

The effect is subtly different if XPath 1.0 compatibility mode is enabled. In this case coercion takes the string value of the element node. This differs from the typed value only in the case where the element has been schema-validated and has a type annotation other than xs:string.

xs:anyURI("urn:dummy")

Supplying an instance of xs:anyURI where the expected type is xs:string is permitted; this is one of the pairs of types where implicit casting is allowed.

17.2

Supplying a number where a string is expected raises a type error.

However, if XPath 1.0 compatibility mode is enabled, the number is converted to a string as if by the fn:string function.

//author/@id

Supplying a sequence of nodes where a single string is expected will raise a type error unless either there is only one node in the sequence. In this case the typed value of the node will be used (this must be of type xs:string, xs:untypedAtomic, or xs:anyURI).

If XPath 1.0 compatibility mode is enabled, however, all strings after the first are discarded, and the string value of the first node is used; if the sequence is empty, a zero-length string is supplied.

("red", "green", "blue")

Supplying a sequence of strings where a single string is expected raises a type error.

If XPath 1.0 compatibility mode is enabled, however, all strings after the first are discarded; the effect is as if the supplied value were "red".

()

Supplying an empty sequence where a single string is expected will fail.

If XPath 1.0 compatibility mode is enabled, however, the value is coerced by applying the function fn:string(()), which delivers the zero-length string.

["a|b"]

Supplying an array holding a single string succeeds, because the rules cause the array to be atomized, and the value after atomization is a single string.

Supplying an array holding multiple strings would fail.

In XPath 1.0 compatibility mode, supplying an array will fail, regardless of the array contents, because the fn:string function does not accept arrays.

 

Example: Coercion to xs:decimal?

Consider the case where the required type (of a variable, or a function argument) is xs:decimal?. For example, the first argument of fn:seconds, which expects a decimal number of seconds. The table below illustrates the values that might be supplied, and the coercions that are applied.

Supplied Value Coercion
12.4

None; the supplied value is an instance of the required type.

()

None; an empty sequence is an instance of the required type.

42

None; the supplied value is an instance of xs:integer, which is a subtype of the required type.

math:pi()

The supplied value is an instance of xs:double, which can be converted to xs:decimal under the coercion rules.

("a", "b")[.="c"]

The supplied value is an empty sequence, which is a valid instance of the required type xs:decimal?. However, the processor may (optionally) reject this as an implausible coercion, on the grounds that it can only succeed in one special case, namely where the filter expression selects no values.

(1.5, 2.5, 3.5)

A type error is raised, except in the case where XPath 1.0 compatibility is enabled, in which case all values after the first are discarded.

<a>3.14159</a>

The element node is atomized; unless it has been schema-validated, the result will be "3.14159" as an instance of xs:untypedAtomic. This is converted to an instance of xs:decimal following the rules of the cast as operator.

"12.2"

Supplying a string where an xs:decimal is a type error, even if XPath 1.0 compatibility mode is enabled. The rules for compatibility mode would allow conversion if the required type were xs:double, but not for xs:decimal .

[1.5]

The array is atomized, and the result is a valid instance of the required type xs:decimal?

[]

The array is atomized, and the result is an empty sequence, which is a valid instance of the required type xs:decimal?

 

Example: Coercion to xs:positive-integer

Consider the case where the required type (of a variable, or a function argument) is xs:positive-integer. The table below illustrates the values that might be supplied, and the coercions that are applied.

Supplied Value Coercion
12

The supplied value is of type xs:integer. Because the supplied value and the required type, xs:positiveInteger, both come under the primitive type xs:decimal, and the value 12 is within the value space of xs:positiveInteger, the value is relabeled as an xs:positiveInteger and the call succeeds.

12.1

This fails with a type error, because the xs:decimal value 12.1 is not a value in the value space of xs:positiveInteger. This is so even though casting to xs:positiveInteger would succeed.

math:pi()

This fails with a type error. A value of type xs:double is accepted where the required type is xs:decimal or xs:float, but not where it is xs:positiveInteger.

<a>1200</a>

The supplied element node is atomized. If the element has not been schema-validated, the result will be an xs:untypedAtomic item, which is successfully cast to the required type xs:positiveInteger. If the element has been validated against a schema, then coercion succeeds if the typed value would itself be acceptable, for example if it is an xs:positiveInteger, or some other xs:decimal within the value space of xs:positiveInteger.

 

Example: Coercion to a union type

Consider the first parameter of the function fn:char, whose declared type is (xs:string | xs:positiveInteger). The rules are the same as if it were a union typed declared in an imported schema.

Supplied Value Coercion
"amp"

The supplied value is of type xs:string, which is one of the allowed types. The call therefore succeeds.

"#"

The supplied value is of type xs:string, which is one of the allowed types. As far as the coercion rules are concerned, the call therefore succeeds. Under the semantic rules for the fn:char function, however, this value is not accepted; a dynamic error (as distinct from a type error) is therefore raised.

0x25

The supplied value is of type xs:integer. Although this is not one of the allowed types, it is acceptable because coercion of the value to type xs:positiveInteger succeeds. The value is relabeled as an instance of xs:positiveInteger.

<a>0x25</a>

The supplied element node is atomized. Assuming that the node has not been schema-validated, the result is an instance of xs:untypedAtomic. The member types of the choice are tested in order. Conversion to xs:string with the value "0x25" succeeds, so the fn:char function is called supplying this string; but the function rejects this string as semantically invalid. The same would happen if the value were, say, <a>37</a>. Supplying such a value requires an explicit cast, for example fn:char( xs:positiveInteger( ./a )).

 

Example: Coercion to a choice type

Suppose the required type is (record(x as xs:decimal, y as xs:decimal, *) | record(size as enum("S", "M", "L", "XL"), *)).

Supplied Value Coercion
{ "x": 1, "y": 2, "z": 3 }

The supplied value is an instance of the first record type: no coercion is necessary.

{ "size": "M" }

The supplied value is an instance of the second record type: no coercion is necessary.

{ "x": 1, "y": 2, "size": "XL" }

The supplied value is an instance of both record types: no coercion is necessary.

{ "x": 1e0, "y": 2e0, "size": "XL" }

The supplied value is not an instance of the first record type because the fields are of type xs:double rather than xs:decimal. It is however an instance of the second record type. It is therefore accepted as is; the fields x and y are not converted from xs:double to xs:decimal.

{ "x": 1e0, "y": 2e0, "size": "XXL" }

The supplied value is not an instance of the first record type because the fields are of type xs:double rather than xs:decimal, and it is not an instance of the second record type because the size value does not match the enumeration type. Coercion is therefore attempted to the first record type, and succeeds. The x and y fields are coerced to xs:decimal, and the size field is accepted as is.

3.5 Schema Types

[Definition: A schema type is a complex type or simple type as defined in the [XML Schema 1.0] or [XML Schema 1.1] specifications, including built-in types as well as user-defined types.]

Every schema type is either a complex type or a simple type; simple types are further subdivided into list types, union types, and atomic types (see [XML Schema 1.0] or [XML Schema 1.1] for definitions and explanations of these terms.)

A schema type can appear as a type annotation on an element or attribute node. The type annotation on an element node can be a complex type or a simple type; the type annotation on an attribute node is always a simple type. Non-instantiable types such as xs:NOTATION or xs:anyAtomicType never appear as type annotations, but their derived types can be so used. Union types never appear as type annotations; when an element or attribute is validated against a union type, the resulting type annotation will be one of the types in the transitive membership of the union type.

[Definition: An atomic type is a simple schema type whose {variety} is atomic.]

An atomic type is either a built-in atomic type (defined either in the XSD specification or in this specification), or it is a user-defined atomic type included in an imported schema.

The in-scope schema types in the static context are initialized with a set of predefined schema types that is determined by the host language. This set may include some or all of the schema types in the namespace http://www.w3.org/2001/XMLSchema, represented in this document by the namespace prefix xs. The schema types in this namespace are defined in [XML Schema 1.0] or [XML Schema 1.1] and augmented by additional types defined in [XQuery and XPath Data Model (XDM) 4.0]. An implementation that has based its type system on [XML Schema 1.0] is not required to support the xs:dateTimeStamp or xs:error types.

The schema types defined in Section 2.8.3 Predefined TypesDM40 are summarized below.

  1. [Definition: xs:untyped is used as the type annotation of an element node that has not been validated, or has been validated in skip mode.] No predefined schema types are derived from xs:untyped.

  2. [Definition: xs:untypedAtomic is an atomic type that is used to denote untyped atomic data, such as text that has not been assigned a more specific type.] An attribute that has been validated in skip mode is represented in the data model by an attribute node with the type annotation xs:untypedAtomic. No predefined schema types are derived from xs:untypedAtomic.

  3. [Definition: xs:dayTimeDuration is derived by restriction from xs:duration. The lexical representation of xs:dayTimeDuration is restricted to contain only day, hour, minute, and second components.]

  4. [Definition: xs:yearMonthDuration is derived by restriction from xs:duration. The lexical representation of xs:yearMonthDuration is restricted to contain only year and month components.]

  5. [Definition: xs:anyAtomicType is an atomic type that includes all atomic items (and no values that are not atomic). Its base type is xs:anySimpleType from which all simple types, including atomic, list, and union types, are derived. All primitive atomic types, such as xs:decimal and xs:string, have xs:anyAtomicType as their base type.]

    Note:

    xs:anyAtomicType will not appear as the type of an actual value in an XDM instance.

  6. [Definition: xs:error is a simple type with no value space. It is defined in Section 3.16.7.3 xs:error XS11-1 and can be used in the 3.1 Sequence Types to raise errors.]

The relationships among the schema types in the xs namespace are illustrated in Figure 2. A more complete description of the XPath 4.0 type hierarchy can be found in Section 1.8 Type SystemFO40.

Type Hierarchy Diagram

Figure 2: Hierarchy of Schema Types used in XPath 4.0.

4 Expressions

This section discusses each of the basic kinds of expression. Each kind of expression has a name such as PathExpr, which is introduced on the left side of the grammar production that defines the expression. Since XPath 4.0 is a composable language, each kind of expression is defined in terms of other expressions whose operators have a higher precedence. In this way, the precedence of operators is represented explicitly in the grammar.

The order in which expressions are discussed in this document does not reflect the order of operator precedence. In general, this document introduces the simplest kinds of expressions first, followed by more complex expressions. For the complete grammar, see Appendix [A XPath 4.0 Grammar].

The highest-level symbol in the XPath grammar is XPath.

XPath ::= Expr
Expr ::= (ExprSingle ++ ",")
ExprSingle ::= ForExpr
| LetExpr
| QuantifiedExpr
| IfExpr
| OrExpr

The XPath 4.0 operator that has lowest precedence is the comma operator, which is used to combine two operands to form a sequence. As shown in the grammar, a general expression (Expr) can consist of multiple ExprSingle operands, separated by commas.

The name ExprSingle denotes an expression that does not contain a top-level comma operator (despite its name, an ExprSingle may evaluate to a sequence containing more than one item.)

The symbol ExprSingle is used in various places in the grammar where an expression is not allowed to contain a top-level comma. For example, each of the arguments of a function call must be a ExprSingle, because commas are used to separate the arguments of a function call.

After the comma, the expressions that have next lowest precedence are ForExpr, LetExpr, QuantifiedExpr, IfExpr, and OrExpr. Each of these expressions is described in a separate section of this document.

4.1 Comments

Comment ::= "(:" (CommentContents | Comment)* ":)"
/* ws: explicit */
/* gn: comments */
CommentContents ::= (Char+ - (Char* ('(:' | ':)') Char*))
/* ws: explicit */

Comments may be used to provide information relevant to programmers who read an expression. Comments are lexical constructs only, and do not affect expression processing.

Comments are strings, delimited by the symbols (: and :). Comments may be nested.

A comment may be used anywhere ignorable whitespace is allowed (see A.3.5.1 Default Whitespace Handling).

The following is an example of a comment:

(: Houston, we have a problem :)

4.2 Primary Expressions

[Definition: Primary expressions are the basic primitives of the language. They include literals, variable references, context value references, and function calls. A primary expression may also be created by enclosing any expression in parentheses, which is sometimes helpful in controlling the precedence of operators.] Map and Array Constructors are described in 4.13.1 Maps and 4.13.2 Arrays.

PrimaryExpr ::= Literal
| VarRef
| ParenthesizedExpr
| ContextValueRef
| FunctionCall
| FunctionItemExpr
| MapConstructor
| ArrayConstructor
| StringTemplate
| UnaryLookup
FunctionItemExpr ::= NamedFunctionRef | InlineFunctionExpr

4.2.1 Literals

Literal ::= NumericLiteral | StringLiteral

[Definition: A literal is a direct syntactic representation of an atomic item.] XPath 4.0 supports two kinds of literals: numeric literals and string literals.

4.2.1.1 Numeric Literals

Changes in 4.0  

  1. Numeric literals can now be written in hexadecimal or binary notation; and underscores can be included for readability.   [Issue 429 PR 433 25 April 2023]

NumericLiteral ::= IntegerLiteral | HexIntegerLiteral | BinaryIntegerLiteral | DecimalLiteral | DoubleLiteral
IntegerLiteral ::= Digits
/* ws: explicit */
HexIntegerLiteral ::= "0x" HexDigits
/* ws: explicit */
BinaryIntegerLiteral ::= "0b" BinaryDigits
/* ws: explicit */
DecimalLiteral ::= ("." Digits) | (Digits "." Digits?)
/* ws: explicit */
DoubleLiteral ::= (("." Digits) | (Digits ("." Digits?)?)) [eE] [+-]? Digits
/* ws: explicit */
Digits ::= DecDigit ((DecDigit | "_")* DecDigit)?
/* ws: explicit */
DecDigit ::= [0-9]
/* ws: explicit */
HexDigits ::= HexDigit ((HexDigit | "_")* HexDigit)?
/* ws: explicit */
HexDigit ::= [0-9a-fA-F]
/* ws: explicit */
BinaryDigits ::= BinaryDigit ((BinaryDigit | "_")* BinaryDigit)?
/* ws: explicit */
BinaryDigit ::= [01]
/* ws: explicit */

The value of a numeric literal is determined as follows (taking the rules in order):

  1. Underscore characters are stripped out. Underscores may be included in a numeric literal to aid readability, but have no effect on the value. For example, 1_000_000 is equivalent to 1000000.

    Note:

    Underscores must not appear at the beginning or end of a sequence of digits, only in intermediate positions. Multiple adjacent underscores are allowed.

  2. A HexIntegerLiteral represents a non-negative integer expressed in hexadecimal: for example 0xffff represents the integer 65535, and 0xFFFF_FFFF represents the integer 4294967295.

  3. A BinaryIntegerLiteral represents a non-negative integer expressed in binary: for example 0b101 represents the integer 5, and 0b1111_1111 represents the integer 255.

  4. The value of a numeric literal containing no . and no e or E character is an atomic item of type xs:integer; the value is obtained by casting from xs:string to xs:integer as specified in Section 21.2 Casting from xs:string and xs:untypedAtomicFO40.

  5. The value of a numeric literal containing . but no e or E character is an atomic item of type xs:decimal; the value is obtained by casting from xs:string to xs:decimal as specified in Section 21.2 Casting from xs:string and xs:untypedAtomicFO40.

  6. The value of a numeric literal containing an e or E character is an atomic item of type xs:double; the value is obtained by casting from xs:string to xs:double as specified in Section 21.2 Casting from xs:string and xs:untypedAtomicFO40.

Note:

The value of a numeric literal is always non-negative. An expression may appear to include a negative number such as -1, but this is technically an arithmetic expression comprising a unary minus operator followed by a numeric literal.

Note:

The effect of the above rules is that in the case of an integer or decimal literal, a dynamic error [err:FOAR0002]FO40 will generally be raised if the literal is outside the range of values supported by the implementation (other options are available: see Section 4.2 Arithmetic operators on numeric valuesFO40 for details.)

The XML Schema specification allows implementations to impose a limit (which must not be less than 18 digits) on the size of integer and decimal values. The full range of values of built-in subtypes of xs:integer, such as xs:long and xs:unsignedLong, can be supported only if the limit is 20 digits or higher. Negative numbers such as the minimum value of xs:long (-9223372036854775808) are technically unary expressions rather than literals, but implementations may prefer to ensure that they are expressible.

Here are some examples of numeric literals:

  • 12 denotes the xs:integer value twelve.

  • 1_000_000 denotes the xs:integer value one million.

  • 12.5 denotes the xs:decimal value twelve and one half.

  • 3.14159_26535_89793e0 is an xs:double value representing the mathematical constant π to 15 decimal places.

  • 125E2 denotes the xs:double value twelve thousand, five hundred.

  • 0xffff denotes the xs:integer value 65535.

  • 0b1000_0001 denotes the xs:integer value 129.

4.2.1.2 String Literals
StringLiteral ::= AposStringLiteral | QuotStringLiteral
/* ws: explicit */
AposStringLiteral ::= "'" (EscapeApos | [^'])* "'"
/* ws: explicit */
QuotStringLiteral ::= '"' (EscapeQuot | [^"])* '"'
/* ws: explicit */
EscapeQuot ::= '""'
/* ws: explicit */
EscapeApos ::= "''"
/* ws: explicit */

The value of a string literal is an atomic item whose type is xs:string and whose value is the string denoted by the characters between the delimiting apostrophes or quotation marks. If the literal is delimited by apostrophes, two adjacent apostrophes within the literal are interpreted as a single apostrophe. Similarly, if the literal is delimited by quotation marks, two adjacent quotation marks within the literal are interpreted as one quotation mark.

Here are some examples of string literals:

  • "He said, ""I don't like it.""" denotes a string containing two quotation marks and one apostrophe.

  • In XQuery, the string literal "&lt;" denotes a string of length 1 containing the single character "<". In XPath, the string literal "&lt;" denotes a string of length 4 containing the four characters "&", "l", "t", ";". (However, when the XPath expression is embedded in an XML document, the sequence "&lt;" will typically have already been converted to "<" by the XML parser.)

Note:

When XPath or XQuery expressions are embedded in contexts where quotation marks have special significance, such as inside XML attributes, or in string literals in a host language such as Java or C#, then additional escaping may be needed.

Note:

Fixed string values can also be written as string templates: see 4.9.2 String Templates. A string template with no enclosed expressions, such as `Jamaica` evaluates to the same value as the string literals "Jamaica" or 'Jamaica'. A string template can contain both single and double quotation marks: `He said: "I don't like it"`. However, there there are some subtle differences:

  • In string literals, the treatment of character and entity references such as &amp; varies between XQuery and XPath; in string templates, such references are not expanded in either language.

  • String templates can only be used where an expression is expected. String literals are also used in some non-expression contexts, for example in defining an enumeration type: see 3.2.6 Enumeration Types.

  • Curly brackets (U+007B (LEFT CURLY BRACKET, {) and U+007D (RIGHT CURLY BRACKET, }) ) and backticks (U+0060 (GRAVE ACCENT, BACKTICK, `) ) have a reserved meaning in string templates.

4.2.1.3 Constants of Other Types

The xs:boolean values true and false can be constructed by calls to the system functions fn:true() and fn:false(), respectively.

Values of other simple types can be constructed by calling the constructor function for the given type. The constructor functions for XML Schema built-in types are defined in Section 20.1 Constructor functions for XML Schema built-in atomic typesFO40. In general, the name of a constructor function for a given type is the same as the name of the type (including its namespace). For example:

  • xs:integer("12") returns the integer value twelve.

  • xs:date("2001-08-25") returns an item whose type is xs:date and whose value represents the date 25th August 2001.

  • xs:dayTimeDuration("PT5H") returns an item whose type is xs:dayTimeDuration and whose value represents a duration of five hours.

Constructor functions can also be used to create special values that have no literal representation, as in the following examples:

  • xs:float("NaN") returns the special floating-point value, "Not a Number."

  • xs:double("INF") returns the special double-precision value, "positive infinity."

Constructor functions are available for all simple types, including union types. For example, if my:dt is a user-defined union type whose member types are xs:date, xs:time, and xs:dateTime, then the expression my:dt("2011-01-10") creates an atomic item of type xs:date. The rules follow XML Schema validation rules for union types: the effect is to choose the first member type that accepts the given string in its lexical space.

It is also possible to construct values of various types by using a cast expression. For example:

  • 9 cast as hatsize returns the atomic item 9 whose type is hatsize.

4.2.2 Variable References

VarRef ::= "$" EQName
EQName ::= QName | URIQualifiedName

[Definition: A variable reference is an EQName preceded by a $-sign.] An unprefixed variable reference is in no namespace. Two variable references are equivalent if their expanded QNames are equal (as defined by the eq operator). The scope of a variable binding is defined separately for each kind of expression that can bind variables.

Every variable reference must match a name in the in-scope variables.

Every variable binding has a static scope. The scope defines where references to the variable can validly occur. It is a static error [err:XPST0008] to reference a variable that is not in scope. If a variable is bound in the static context for an expression, that variable is in scope for the entire expression except where it is occluded by another binding that uses the same name within that scope.

At evaluation time, the value of a variable reference is the value to which the relevant variable is bound.

4.2.3 Context Value References

ContextValueRef ::= "."

A context value reference evaluates to the context value.

In many syntactic contexts, the context value will be a single item. For example this applies on the right-hand side of the / or ! operators, or within a Predicate.

If the context value is absentDM40, a context value reference raises a type error [err:XPDY0002].

Note:

Being absent is not the same thing as being empty.

4.2.4 Parenthesized Expressions

ParenthesizedExpr ::= "(" Expr? ")"

Parentheses may be used to override the precedence rules. For example, the expression (2 + 4) * 5 evaluates to thirty, since the parenthesized expression (2 + 4) is evaluated first and its result is multiplied by five. Without parentheses, the expression 2 + 4 * 5 evaluates to twenty-two, because the multiplication operator has higher precedence than the addition operator.

Empty parentheses are used to denote an empty sequence, as described in 4.7.1 Sequence Concatenation.

4.2.5 Enclosed Expressions

EnclosedExpr ::= "{" Expr? "}"

[Definition: An enclosed expression is an instance of the EnclosedExpr production, which allows an optional expression within curly brackets.] [Definition: In an enclosed expression, the optional expression enclosed in curly brackets is called the content expression.] If the content expression is not provided explicitly, the content expression is ().

Note:

Despite the name, an enclosed expression is not actually an expression in its own right; rather it is a construct that is used in the grammar of many other expressions.

4.3 Postfix Expressions

PostfixExpr ::= PrimaryExpr | FilterExpr | DynamicFunctionCall | LookupExpr | FilterExprAM
FilterExpr ::= PostfixExpr Predicate
DynamicFunctionCall ::= PostfixExpr PositionalArgumentList
LookupExpr ::= PostfixExpr Lookup

A postfix expression takes one of the following forms:

  • [Definition: A filter expression is an expression in the form E1[E2]: its effect is to return those items from the value of E1 that satisfy the predicate in E2.]

    Filter expressions are described in 4.4 Filter Expressions.

    An example of a filter expression is (1 to 100)[. mod 2 = 0] which returns all even numbers in the range 1 to 100.

    The base expression E1 can itself be a postfix expression, so multiple predicates are allowed, in the form E1[E2][E3][E4].

  • An expression (other than a raw EQName) followed by an argument list in parentheses (that is, E1(E2, E3, ...)) is referred to as a dynamic function call. Its effect is to evaluate E1 to obtain a function, and then call that function, with E2, E3, ... as arguments. Dynamic function calls are described in 4.5.2.1 Dynamic Function Calls.

    An example of a dynamic function call is $f("a", 2) where the value of variable $f must be a function item.

  • A lookup-expression takes the form E1?K, where E1 is an expression returning a sequence of maps or arrays, and K is a key specifier, which indicates which entries in a map, or members in an array, should be selected.

    Lookup expressions are described in 4.13.3.1 Postfix Lookup Expressions.

    An example of a lookup expression is $emp?name, where the value of variable $emp is a map, and the string "name" is the key of one of the entries in the map.

Postfix expressions are evaluated from left-to-right. For example, the expression $E1[E2]?(E3)(E4) is evaluated by first evaluating the filter expression $E1[E2] to produce a sequence of maps and arrays (say $S), then evaluating the lookup expression $S?(E3) to produce a function item (say $F), then evaluating the dynamic function call $F(E4) to produce the final result.

Note:

The grammar for postfix expressions is defined here in a way designed to link clearly to the semantics of the different kinds of expression. For parsing purposes, the equivalent production rule:

PostfixExpr := PrimaryExpr (Predicate | PositionalArgumentList | Lookup)*

(as used in XPath 3.1) is probably more convenient.

4.4 Filter Expressions

Changes in 4.0  

  1. The value of a predicate in a filter expression can now be a sequence of integers.   [Issue 816 PR 996 6 February 2024]

FilterExpr ::= PostfixExpr Predicate
Predicate ::= "[" Expr "]"

A filter expression consists of a base expression followed by a predicate, which is an expression written in square brackets. The result of the filter expression consists of the items returned by the base expression, filtered by applying the predicate to each item in turn. The ordering of the items returned by a filter expression is the same as their order in the result of the primary expression.

Note:

Where the expression before the square brackets is a ReverseStep or ForwardStep, the expression is technically not a filter expression but an AxisStep. There are minor differences in the semantics: see 4.6.5 Predicates within Steps

Here are some examples of filter expressions:

  • Given a sequence of products in a variable, return only those products whose price is greater than 100.

    $products[price gt 100]
  • List all the integers from 1 to 100 that are divisible by 5. (See 4.7.1 Sequence Concatenation for an explanation of the to operator.)

    (1 to 100)[. mod 5 eq 0]
  • The result of the following expression is the integer 25:

    (21 to 29)[5]
  • The following example returns the fifth through ninth items in the sequence bound to variable $orders.

    $orders[5 to 9]
  • The following example illustrates the use of a filter expression as a step in a path expression. It returns the last chapter or appendix within the book bound to variable $book:

    $book/(chapter | appendix)[last()]

For each item in the input sequence, the predicate expression is evaluated using an inner focus, defined as follows: The context value is the item currently being tested against the predicate. The context size is the number of items in the input sequence. The context position is the position of the context value within the input sequence.

For each item in the input sequence, the result of the predicate expression is coerced to an xs:boolean value, called the predicate truth value, as described below. Those items for which the predicate truth value is true are retained, and those for which the predicate truth value is false are discarded.

[Definition: The predicate truth value of a value $V is the result of the expression if ($V instance of xs:numeric+) then ($V = position()) else fn:boolean($V).]

Expanding this definition, the predicate truth value can be obtained by applying the following rules, in order:

  1. If the value V of the predicate expression is a sequence whose first item is an instance of the type xs:numeric, then:

    1. V must be an instance of the type xs:numeric+ (that is, every item in V must be numeric). A type error [err:FORG0006]FO40 is raised if this is not the case.

    2. The predicate truth value is true if V is equal (by the = operator) to the context position, and is false otherwise.

    In effect this means that an item in the input sequence is selected if its position in the sequence is equal to one or more of the numeric values in the predicate. For example, the predicate [3 to 5] is true for the third, fourth, and fifth items in the input sequence.

    Note:

    It is possible, though not generally useful, for the value of a numeric predicate to depend on the focus, and thus to differ for different items in the input sequence. For example, the predicate [xs:integer(@seq)] selects those items in the input sequence whose @seq attribute is numerically equal to their position in the input sequence.

    It is also possible, and again not generally useful, for the value of the predicate to be numeric for some items in the input sequence, and boolean for others. For example, the predicate [@special otherwise last()] is true for an item that either has an @special attribute, or is the last item in the input sequence.

    Note:

    The truth value of a numeric predicate does not depend on the order of the numbers in V. The predicates [ 1, 2, 3 ] and [ 3, 2, 1 ] have exactly the same effect. The items in the result of a filter expression always retain the ordering of the input sequence.

    Note:

    The truth value of a numeric predicate whose value is non-integral or non-positive is always false.

    Note:

    Beware that using boolean operators (and, or, not()) with numeric values may not have the intended effect. For example the predicate [1 or last()] selects every item in the sequence, because or operates on the effective boolean value of its operands. The required effect can be achieved with the predicate [1, last()].

  2. Otherwise, the predicate truth value is the effective boolean value of the predicate expression.

4.5 Functions

Functions in XPath 4.0 arise in two ways:

The functions defined by a statically known function definition can be invoked using a static function call. Function items corresponding to these definitions can also be obtained, as dynamic values, by evaluating a named function reference. Function items can also be obtained using the fn:function-lookup function: in this case the function name and arity do not need to be known statically, and the function definition need not be present in the static context, so long as it is in the dynamic context.

Static and dynamic function calls are described in the following sections.

4.5.1 Static Function Calls

The static context for an expression includes a set of statically known function definitions. Every function definition in the static context has a name (which is an expanded QName) and an arity range, which is a range of permitted arities for calls on that function. Two function definitions having the same name must not have overlapping arity ranges. This means that for a given static function call, it is possible to identify the target function definition in the static context unambiguously from knowledge of the function name and the number of supplied arguments.

A static function call is bound to a function definition in the static context by matching the name and arity. If the function call has P positional arguments followed by K keyword arguments, then the required arity is P+K, and the static context must include a function definition whose name matches the expanded QName in the function call, and whose arity range includes this required arity. This is the function chosen to be called. The result of the function is obtained by evaluating the expression that forms its implementation, with a dynamic context that provides values for all the declared parameters, initialized as described in 4.5.1.2 Evaluating Static Function Calls below.

Similarly, a function reference of the form f#N binds to a function definition in the static context whose name matches f where MinP ≤ N and MaxP ≥ N. The result of evaluating a function reference is a function item which can be called using a dynamic function call. Function items are never variadic and their arguments are always supplied positionally. For example, the function reference fn:concat#3 returns a function item with arity 3, which is always called by supplying three positional arguments, and whose effect is the same as a static call on fn:concat with three positional arguments.

The detailed rules for evaluating static function calls and function references are defined in subsequent sections.

4.5.1.1 Static Function Call Syntax

Changes in 4.0  

  1. Keyword arguments are allowed on static function calls, as well as positional arguments.   [Issue 155 PR 159 30 September 2020]

FunctionCall ::= EQName ArgumentList
/* xgc: reserved-function-names */
/* gn: parens */
ArgumentList ::= "(" ((PositionalArguments ("," KeywordArguments)?) | KeywordArguments)? ")"
PositionalArguments ::= (Argument ++ ",")
Argument ::= ExprSingle | ArgumentPlaceholder
ArgumentPlaceholder ::= "?"
KeywordArguments ::= (KeywordArgument ++ ",")
KeywordArgument ::= EQName ":=" Argument

[Definition: A static function call consists of an EQName followed by a parenthesized list of zero or more arguments.]

The argument list consists of zero or more positional arguments, followed by zero or more keyword arguments.

[Definition: An argument to a function call is either an argument expression or an ArgumentPlaceholder (?); in both cases it may either be supplied positionally, or identified by a name (called a keyword).]

This section is concerned with static function calls in which none of the arguments are ArgumentPlaceholders. Calls using one or more ArgumentPlaceholders are covered in the section 4.5.2.3 Partial Function Application.

If the function name supplied in a static function call is an unprefixed lexical QName, it is expanded using the default function namespace in the static context.

In other cases, The expanded QName used as the function name and the number of arguments used in the static function call (the required arity) must match the name and arity range of a function definition in the static context using the rules defined in the previous section; if there is no match, a static error is raised [err:XPST0017].

Evaluation of static function calls is described in 4.5.1.2 Evaluating Static Function Calls , while evaluation of dynamic function calls is described in 4.5.2.2 Evaluating Dynamic Function Calls .

Since the arguments of a function call are separated by commas, any argument expression that contains a top-level comma operator must be enclosed in parentheses. Here are some illustrative examples of static function calls:

  • my:three-argument-function(1, 2, 3) denotes a static function call with three positional arguments. The corresponding function declaration must define at least three parameters, and may define more, provided they are optional.

  • my:two-argument-function((1, 2), 3) denotes a static function call with two arguments, the first of which is a sequence of two values. The corresponding function declaration must define at least two parameters, and may define more, provided they are optional.

  • my:two-argument-function(1, ()) denotes a static function call with two arguments, the second of which is an empty sequence.

  • my:one-argument-function((1, 2, 3)) denotes a static function call with one argument that is a sequence of three values.

  • my:one-argument-function(( )) denotes a static function call with one argument that is an empty sequence.

  • my:zero-argument-function( ) denotes a static function call with zero arguments.

  • lang(node := $n, language := 'de') is a static function call with two keyword arguments. The corresponding function declaration defines two parameters, a required parameter language and an optional parameter node. This call supplies values for both parameters. It is equivalent to the call fn:lang('de', $n). Note that the keyword arguments are in a different order from the parameter declarations.

  • sort(//employee, key := fn($e) { xs:decimal($e/salary) }) is a static function call with one positional argument and one keyword argument. The corresponding function declaration defines three parameters, a required parameter $input, an optional parameter $collation, and an optional parameter $key This call supplies values for the first and third parameters, leaving the second parameter ($collation) to take its default value. The default value of the $collation parameter is given as fn:default-collation(), so the value supplied to the function is the default collation from the dynamic context of the caller. It is equivalent to the call fn:sort(//employee, fn:default-collation(), fn($e) { xs:decimal($e/salary) }).

An EQName in a KeywordArgument is expanded to a QName value; if there is no prefix, then the name is in no namespace (otherwise the prefix is resolved in the usual way). The keywords used in a function call (after expansion to QNames) must be distinct [err:XPST0017]; [err:XPST0017].

4.5.1.2 Evaluating Static Function Calls

This section applies to static function calls where none of the arguments is an ArgumentPlaceholder. For function calls involving placeholders, see 4.5.2.3 Partial Function Application.

When a static function call FC is evaluated with respect to a static context SC and a dynamic context DC, the result is obtained as follows:

  1. The function definition FD to be used is found in the statically known function definitions of SC.

    The required arity is the total number of arguments in the function call, including both positional and keyword arguments.

    There can be at most one function definition FD in the statically known function definitions component of SC whose function name matches the expanded QName in FC and whose arity range includes the arity of FC’s ArgumentList.

    If there is no such function definition, a static error [err:XPST0017] is raised.

  2. Each parameter in the function definition FD is matched to an argument expression as follows:

    1. If there are N positional arguments in the function call FC, then the corresponding argument expressions are matched pairwise to the first N parameters in the declaration. For this purpose the required parameters and optional parameters in FD are concatenated into a single list, in order.

    2. Any keyword arguments in FC are then matched to parameters (whether required or optional) in FD by comparing the keyword used in FC with the paramater name declared in FD. Each keyword must match the name of a declared parameter [err:XPST0017], and this must be one that has not already been matched to a positional argument. [err:XPST0017].

    3. If any required parameter has not been matched to any argument in FC by applying the above rules, a static error is reported [err:XPST0017]

    4. If any optional parameter has not been matched to any argument in FC by applying the above rules, then the parameter is matched to the default value expression for that parameter in FD.

  3. Each argument expression established by the above rules is evaluated with respect to DC. The order of argument evaluation is implementation dependent and it is not required that an argument be evaluated if the function body can be evaluated without evaluating that argument.

    Note:

    All argument expressions, including default value expressions, are evaluated in the dynamic context of the function call. It is therefore possible to use a default value expression such as ., or /, or fn:current-dateTime(), whose value depends on the dynamic context of the function call.

    If the expression used for the default value of a parameter has no dependencies on the dynamic context, then an implementation may choose to reuse the same value on repeated function calls rather than re-evaluating it on each function call.

    Note:

    This is relevant, for example, if the expression constructs new nodes.

  4. The result of evaluating the argument expression is converted to the required type (the declared type associated with the corresponding parameter in the function declaration, defaulting to item()*) by applying the coercion rules.

    This applies both to explicitly supplied arguments, and to values obtained by evaluating default value expressions. In both cases a type error will be raised if the value (after coercion) does not match the required type.

  5. The result of the function call is obtained as follows:

    • FD’s body is invoked in an implementation-dependent way. The processor makes the following information available to that invocation:

      • The converted argument values;

      • An empty set of nonlocal variable bindings; and

      • If the function is context dependent, the static context SC and dynamic context DC of the function call.

      How this information is used is implementation-defined. An API used to call external functions must state how the static and dynamic contexts are provided to a function that is called. The F&O specification states how the static and dynamic contexts are used in each function that it defines. A host language must state how the static and dynamic contexts are used in functions that it provides.

    • The result is converted to the required type (the declared return type in the function declaration, defaulting to item()*) by applying the coercion rules.

      The result of applying the coercion rules is either an instance of FD’s return type or a dynamic error. This result is then the result of evaluating FC.

      Note:

      A host language may define alternative rules for processing the result, especially in the case of external functions implemented using a non-XDM type system.

    • Errors raised by system functions are defined in [XQuery and XPath Functions and Operators 4.0].

    • Errors raised by host-language-dependent functions are implementation-defined.

    Example: A System Function

    The following function call uses the function Section 2.1.5 fn:base-uriFO40. Use of SC and DC and errors raised by this function are all defined in [XQuery and XPath Functions and Operators 4.0].

    base-uri()

4.5.2 Function Items

A function item is an XDM value that can be bound to a variable, or manipulated in various ways by XPath 4.0 expressions. The most significant such expression is a dynamic function call, which supplies values of arguments and evaluates the function to produce a result.

The syntax of dynamic function calls is defined in 4.5.2.1 Dynamic Function Calls.

A number of constructs can be used to produce a function item, notably:

  • A named function reference (see 4.5.2.4 Named Function References) constructs a function item by reference to function definitions in the static context. For example, fn:node-name#1 returns a function item whose effect is to call the static fn:node-name function with one argument.

  • An inline function (see 4.5.2.5 Inline Function Expressions ) constructs a function item whose body is defined locally. For example, the construct fn($x) { $x + 1 } returns a function item whose effect is to increment the value of the supplied argument.

  • A partial function application (see 4.5.2.3 Partial Function Application) derives one function item from another by supplying the values of some of its arguments. For example, fn:ends-with(?, ".txt") returns a function item with one argument that tests whether the supplied string ends with the substring ".txt".

  • Maps and arrays are also function items. See 4.13.1.1 Map Constructors and 4.13.2.1 Array Constructors.

  • The fn:function-lookup function can be called to discover functions that are present in the dynamic context.

  • The fn:load-xquery-module function can be called to load functions dynamically from an external XQuery library module.

  • Some system functions such as fn:random-number-generator and fn:op return a function item as their result.

These constructs are described in detail in the following sections, or in [XQuery and XPath Functions and Operators 4.0].

4.5.2.1 Dynamic Function Calls
DynamicFunctionCall ::= PostfixExpr PositionalArgumentList
PositionalArgumentList ::= "(" PositionalArguments? ")"
PositionalArguments ::= (Argument ++ ",")
Argument ::= ExprSingle | ArgumentPlaceholder
ArgumentPlaceholder ::= "?"

[Definition: A dynamic function call consists of a base expression that returns the function and a parenthesized list of zero or more arguments (argument expressions or ArgumentPlaceholders).]

A dynamic function call is evaluated as described in 4.5.2.2 Evaluating Dynamic Function Calls.

The following are examples of some dynamic function calls:

  • This example calls the function contained in $f, passing the arguments 2 and 3:

    $f(2, 3)
  • This example fetches the second item from sequence $f, treats it as a function and calls it, passing an xs:string argument:

    $f[2]("Hi there")
  • This example calls the function $f passing no arguments, and filters the result with a positional predicate:

    $f()[2]

Note:

Arguments in a dynamic function call are always supplied positionally.

4.5.2.2 Evaluating Dynamic Function Calls

This section applies to dynamic function calls whose arguments do not include an ArgumentPlaceholder. For function calls that include a placeholder, see 4.5.2.3 Partial Function Application.

[Definition: A dynamic function call is an expression that is evaluated by calling a function item, which is typically obtained dynamically.]

When a dynamic function call FC is evaluated, the result is obtained as follows:

  1. The function item FI to be called is obtained by evaluating the base expression of the function call. If this yields a sequence consisting of a single function item whose arity matches the number of arguments in the ArgumentList, let FI denote that function item. Otherwise, a type error is raised [err:XPTY0004].

    Note:

    Keyword arguments are not allowed in a dynamic function call.

  2. Argument expressions are evaluated, producing argument values. The order of argument evaluation is implementation-dependent and an argument need not be evaluated if the function body can be evaluated without evaluating that argument.

  3. Each argument value is converted to the corresponding parameter type in FI’s signature by applying the coercion rules, resulting in a converted argument value

  4. If FI is a map, it is evaluated as described in 4.13.1.2 Maps as Functions.

  5. If FI is an array, it is evaluated as described in 4.13.2.2 Arrays as Functions.

  6. If FI’s body is an XPath 4.0 expression (for example, if FI is an anonymous function, or a partial application of such a function):

    1. FI’s body is evaluated. The static context for this evaluation is the static context of the XPath 4.0 expression. The dynamic context for this evaluation is obtained by taking the dynamic context of the InlineFunctionExpr that contains the FunctionBody, and making the following changes:

      • The focus (context value, context position, and context size) is absentDM40.

      • In the variable values component of the dynamic context, each converted argument value is bound to the corresponding parameter name.

        When this is done, the converted argument values retain their dynamic types, even where these are subtypes of the declared parameter types. For example, a function with a parameter $p of type xs:decimal can be called with an argument of type xs:integer, which is derived from xs:decimal. During the processing of this function call, the value of $p inside the body of the function retains its dynamic type of xs:integer.

      • FI’s nonlocal variable bindings are also added to the variable values. (Note that the names of the nonlocal variables are by definition disjoint from the parameter names, so there can be no conflict.)

    2. The value returned by evaluating the function body is then converted to the declared return type of FI by applying the coercion rules. The result is then the result of evaluating FC.

      As with argument values, the value returned by a function retains its dynamic type, which may be a subtype of the declared return type of FI. For example, a function that has a declared return type of xs:decimal may in fact return a value of dynamic type xs:integer.

    Example: Derived Types and Nonlocal Variable Bindings

    $incr is a nonlocal variable that is available within the function because its variable binding has been added to the variable values of the function.. Even though the parameter and return type of this function are both xs:decimal, the more specific type xs:integer is preserved in both cases.

    let $incr := 1
    let $f := function($i as xs:decimal) as xs:decimal { $i + $incr }
    return $f(5)

     

    Example: Using the Context Value in an Anonymous Function

    The following example will raise a type error [err:XPDY0002]:

    let $vat := function() { @vat + @price }
    return doc('wares.xml')/shop/article/$vat()

    Instead, the context value can be used as an argument to the anonymous function:

    let $vat := function($art) { $art/@vat + $art/@price }
    return doc('wares.xml')/shop/article/$vat(.)

    Alternatively, the value can be referenced as a nonlocal variable binding:

    let $ctx := doc('wares.xml')/shop/article
    let $vat := function() { for $a in $ctx return $a/@vat + $a/@price }
    return $vat()

    Finally, a focus function can be used. This binds the value of the argument to the context value within the function body:

    let $vat := function { @vat + @price }
    return $vat(doc('wares.xml')/shop/article)
  7. If the implementation of FI is not an XPath 4.0 expression (for example, FI is a system function the body of the function is evaluated, and the result is converted to the declared return type, in the same way as for a static function call (see 4.5.1.1 Static Function Call Syntax).

    Errors may be raised in the same way.

4.5.2.3 Partial Function Application

[Definition: A static or dynamic function call is a partial function application if one or more arguments is an ArgumentPlaceholder.]

The rules for partial function application in static function calls and dynamic function calls have a great deal in common, but they are stated separately below for clarity.

In each case, the result of a partial function application is a function item, whose arity is equal to the number of placeholders in the call.

More specifically, the result of the partial function application is a partially applied function. [Definition: A partially applied function is a function created by partial function application.]

For static function calls, the result is obtained as follows:

  1. The function definition FD to be partially applied is determined in the same way as for a static function call without placeholders, as described in 4.5.1.1 Static Function Call Syntax. For this purpose an ArgumentPlaceholder contributes to the count of arguments.

  2. The parameters of FD are classified into three categories:

    • Parameters that map to a placeholder, referred to as placeholder parameters.

    • Parameters for which an explicit value is given in the function call (either positionally or by keyword), referred to as explicitly supplied parameters.

    • Parameters (which are necessarily optional parameters) for which no corresponding argument is supplied, either as a placeholder or with an explicit value. These are referred to as defaulted parameters.

    Note:

    A partial function application need not have any explicitly supplied parameters. For example, the partial function application fn:string(?) is allowed; it has exactly the same effect as the named function reference fn:string#1.

  3. Explicitly supplied parameters and defaulted parameters are evaluated and converted to the required type using the rules for a static function call. This may result in an error being raised.

    A type error is raised if any of the explicitly supplied or defaulted parameters, after applying the coercion rules, does not match the required type of the corresponding parameter.

    In addition, a dynamic error may be raised if any of the explicitly supplied or defaulted parameters does not match other constraints on the value of that parameter (for example, if the value supplied for a parameter expecting a regular expression is not a valid regular expression); or if the processor is able to establish that evaluation of the resulting function will fail for any other reason (for example, if an error is raised while evaluating a subexpression in the function body that depends only on explicitly supplied and defaulted parameters).

    In all cases the error code is the same as for a static function call supplying the same invalid value(s).

    In the particular case where all the supplied arguments are placeholders, the error behavior should be the same as for an equivalent named function reference: for example, fn:id#1 fails if there is no context node, and fn:id(?) should fail likewise.

  4. The result is a partially applied function having the following properties (which are defined in Section 2.9.4 Function ItemsDM40):

    • name: The name of FD if all parameters map to placeholders, that is, if the partial function application is equivalent to the corresponding named function reference. Otherwise, the name is absent.

    • identity: A new function identity distinct from the identity of any other function item.

      Note:

      See also 4.5.2.7 Function Identity.

    • arity: The number of placeholders in the function call.

    • parameter names: The names of the parameters of FD that have been identified as placeholder parameters, retaining the order in which the placeholders appear in the function call.

      Note:

      A partial function application can be used to change the order of parameters, for example fn:contains(substring := ?, value := ?) returns a function item that is equivalent to fn:contains#2, but with the order of arguments reversed.

    • signature: The parameters in the returned function are the parameters of FD that have been identified as placeholder parameters, retaining the order in which the placeholders appear in the function call. The result type of the returned function is the same as the result type of FD.

      An implementation which can determine a more specific signature (for example, through use of type analysis) is permitted to do so.

    • body: The body of FD.

    • captured context: The static and dynamic context of the function call, augmented, for each explicitly supplied parameter and each defaulted parameter, with a binding of the converted argument value to the corresponding parameter name.

    Example: Partial Application of a System Function

    The following partial function application creates a function item that computes the sum of squares of a sequence.

    let $sum-of-squares := fold-right(?, 0, function($a, $b) { $a*$a + $b })
    return $sum-of-squares(1 to 3)

    $sum-of-squares is an anonymous function. It has one parameter, named $seq, which is taken from the corresponding parameter in fn:fold-right (the other two parameters are fixed). The implementation is the implementation of fn:fold-right, which is a context-independent system function. The nonlocal bindings contain the fixed bindings for the second and third parameters of fn:fold-right.

For dynamic function calls, the result is obtained as follows:

  1. The function item FI to be partially applied is determined in the same way as for a dynamic function call without placeholders, as described in 4.5.2.1 Dynamic Function Calls. For this purpose an ArgumentPlaceholder contributes to the count of arguments.

  2. The parameters of FI are classified into two categories:

    • Parameters that map to a placeholder, referred to as placeholder parameters.

    • Parameters for which an explicit value is given in the function call, referred to as supplied parameters.

    Note:

    A partial function application need not have any explicitly supplied parameters. For example, if $f is a function with arity 2, then the partial function application $f(?, ?) returns a function that has exactly the same effect as $f.

  3. Arguments corresponding to supplied parameters are evaluated and converted to the required type of the parameter, using the rules for dynamic function calls.

    A type error is raised if any of the supplied parameters, after applying the coercion rules, does not match the required type.

    In addition, a dynamic error may be raised if any of the supplied parameters does not match other constraints on the value of that parameter (for example, if the value supplied for a parameter expecting a regular expression is not a valid regular expression); or if the processor is able to establish that evaluation of the resulting function will fail for any other reason (for example, if an error is raised while evaluating a subexpression in the function body that depends only on explicitly supplied parameters).

    In both cases the error code is the same as for a dynamic function call supplying the same invalid value.

  4. The result of the partial function application is a partially applied function with the following properties (which are defined in Section 2.9.4 Function ItemsDM40):

    • name: Absent.

    • arity: The number of placeholders in the function call.

    • parameter names: The names of parameters in FI that have been identified as placeholder parameters, in order.

      Note:

      In a dynamic partial function application, argument keywords are not available, so it is not possible to change the order of parameters.

    • signature: The signature of FI, removing the types of supplied parameters. An implementation which can determine a more specific signature (for example, through use of type analysis) is permitted to do so.

    • body: The body of FI.

    • captured context: the captured context of FI, augmented, for each supplied parameter, with a binding of the converted argument value to the corresponding parameter name.

    Example: Partial Application of an Anonymous Function

    In the following example, $f is an anonymous function, and $paf is a partially applied function created from $f.

    let $f := function($seq, $delim) { fold-left($seq, "", concat(?, $delim, ?)) }
    let $paf := $f(?, ".")
    return $paf(1 to 5)

    $paf is also an anonymous function. It has one parameter, named $delim, which is taken from the corresponding parameter in $f (the other parameter is fixed). The implementation of $paf is the implementation of $f, which is fn:fold-left($seq, "", fn:concat(?, $delim, ?)). This implementation is associated with the SC and DC of the original expression in $f. The nonlocal bindings associate the value "." with the parameter $delim.

Partial function application never returns a map or an array. If $f is a map or an array, then $f(?) is a partial function application that returns a function, but the function it returns is neither a map nor an array.

Example: Partial Application of a Map

The following partial function application converts a map to an equivalent function that is not a map.

let $a := { "A": 1, "B": 2 }(?)
return $a("A")
4.5.2.4 Named Function References
NamedFunctionRef ::= EQName "#" IntegerLiteral
/* xgc: reserved-function-names */
EQName ::= QName | URIQualifiedName

[Definition: A named function reference is an expression (written name#arity) which evaluates to a function item, the details of the function item being based on the properties of a function definition in the static context .]

The name and arity of the required function are known statically.

If the EQName is a lexical QName, it is expanded using the default function namespace in the static context.

The expanded QName and arity must correspond to a function definition present in the static context. More specifically, for a named function reference F#N, there must be a function definition in the statically known function definitions whose name matches F, and whose arity range includes N . Call this function definition FD.

If FD is context dependent for the given arity, then the returned function item has a captured context comprising the static and dynamic context of the named function reference.

Note:

In practice, it is necessary to retain only those parts of the static and dynamic context that can affect the outcome. These means it is unnecessary to retain parts of the context that no system function depends on (for example, local variables), or parts that are invariant within an execution scope (for example, the implicit timezone).

Example: A Context-Dependent Named Function Reference

Consider:

let $f := <foo/>/fn:name#0 return <bar/>/$f()

The function fn:name(), with no arguments, returns the name of the context node. The function item delivered by evaluating the expression fn:name#0 returns the name of the element that was the context node at the point where the function reference was evaluated (that is, the <foo> element). This expression therefore returns "foo", not "bar".

An error is raised if the identified function depends on components of the static or dynamic context that are not present, or that have unsuitable values. For example [err:XPDY0002] is raised for the expression fn:name#0 if the context item is absent, and [err:FODC0001]FO is raised for the call fn:id#1 if the context item is not a node in a tree that is rooted at a document node. The error that is raised is the same as the error that would be raised by the corresponding function if called with the same static and dynamic context.

If the expanded QName and arity in a named function reference do not match the name and arity range of a function definition in the static context, a static error is raised [err:XPST0017].

The value of a NamedFunctionRef is a function item FI obtained from FD as follows:

  • name: The name of FD.

  • identity:

    • If FD is context dependent for the given arity, then a new function identity distinct from the identity of any other function item.

      Note:

      In the general case, a function reference to a context-dependent function will produce different results every time it is evaluated, because the resulting function item has a captured context (see Section 2.9.4 Function ItemsDM40) that includes the dynamic context of the particular evaluation. Optimizers, however, are allowed to detect cases where the captured context happens to be the same, or where any variations are immaterial, and where it is therefore safe to return the same function item each time. This might be the case, for example, where the only context dependency of a function is on the default collation, and the default collation for both evaluations is known to be the same.

    • Otherwise, a function identity that is the same as that produced by the evaluation of any other named function reference with the same function name and arity.

      This rule applies even across different execution scopesFO40: for example if a parameter to a call to fn:transform is set to the result of the expression fn:abs#1, then the function item passed as the parameter value will be identical to that obtained by evaluating the expression fn:abs#1 within the target XSLT stylesheet.

      This rule also applies when the target function definition is nondeterministicFO40. For example all evaluations of the named function reference map:keys#2 return identical function items, even though two evaluations of map:keys with the same arguments may produce different results.

    Note:

    See also 4.5.2.7 Function Identity.

  • arity: As specified in the named function reference.

  • parameter names: The first A parameter names of FD, where A is the required arity.

  • signature: Formed from the required types of the first A parameters of FD, and the function result type of FD.

  • body: The body of FD.

  • captured context: Comprises the static and dynamic context of the named function reference, augmented with bindings of the names of parameters of FD beyond the A’th parameter, to their respective default values.

    Note:

    In practice, it is necessary to retain only the parts of the context that the function actually depends on (if any).

Note:

Consider the system function fn:format-date, which has an arity range of 2 to 5. The named function reference fn:format-date#3 returns a function item whose three parameters correspond to the first three parameters of fn:format-date; the remaining two arguments will take their default values. To obtain an arity-3 function that binds to arguments 1, 2, and 5 of fn:format-date, use the partial function application format-date(?, ?, place := ?).

The following are examples of named function references:

  • fn:abs#1 references the fn:abs function which takes a single argument.

  • fn:concat#5 references the fn:concat function which takes 5 arguments.

  • local:myfunc#2 references a function named local:myfunc which takes 2 arguments.

Note:

Function items, as values in the data model, have a fixed arity, and a dynamic function call always supplies the arguments positionally. In effect, the result of evaluating my:func#3 is the same as the result of evaluating the inline function expression fn($x, $y, $z) { my:func($x, $y, $z) }, except that the returned function has a name (it retains the name my:func).

4.5.2.5 Inline Function Expressions

Changes in 4.0  

  1. In inline function expressions, the keyword function may be abbreviated as fn.   [Issue 1192 PR 1197 21 May 2024]

  2. New abbreviated syntax is introduced (focus function) for simple inline functions taking a single argument. An example is fn { ../@code }   [Issue 503 PR 521 30 May 2023]

InlineFunctionExpr ::= ("function" | "fn") FunctionSignature? FunctionBody
FunctionSignature ::= "(" ParamList ")" TypeDeclaration?
ParamList ::= (VarNameAndType ** ",")
VarNameAndType ::= "$" EQName TypeDeclaration?
TypeDeclaration ::= "as" SequenceType
FunctionBody ::= EnclosedExpr

[Definition: An inline function expression , when evaluated, creates an anonymous function defined directly in the inline function expression.] An inline function expression specifies the names and SequenceTypes of the parameters to the function, the SequenceType of the result, and the body of the function.

An inline function expression whose FunctionSignature is omitted is known as a focus function. Focus functions are described in 4.5.2.6 Focus Functions.

[Definition: An anonymous function is a function item with no name. Anonymous functions may be created, for example, by evaluating an inline function expression or by partial function application.]

The keywords function and fn are synonymous.

The syntax allows the names and types of the function argument to be declared, along with the type of the result:

function($x as xs:integer, $y as xs:integer) as xs:integer { $x + $y }

The types can be omitted, and the keyword can be abbreviated:

fn($x, $y) { $x + $y }

A zero-arity function can be written as, for example, fn() { current-date() }.

If a function parameter is declared using a name but no type, its default type is item()*. If the result type is omitted, its default result type is item()*.

The parameters of an inline function expression are considered to be variables whose scope is the function body. It is a static error [err:XQST0039] for an inline function expression to have more than one parameter with the same name.

The static context for the function body is inherited from the location of the inline function expression, with the exception of the static type of the context value which is initially absentDM40.

The variables in scope for the function body include all variables representing the function parameters, as well as all variables that are in scope for the inline function expression.

Note:

Function parameter names can mask variables that would otherwise be in scope for the function body.

The result of an inline function expression is a single function item with the following properties (as defined in Section 2.9.4 Function ItemsDM40):

  • name: Absent.

  • identity: A new function identity distinct from the identity of any other function item.

    Note:

    See also 4.5.2.7 Function Identity.

  • parameter names: The parameter names in the InlineFunctionExpr’s ParamList.

  • signature: A FunctionType constructed from the SequenceTypes in the InlineFunctionExpr. An implementation which can determine a more specific signature (for example, through use of type analysis of the function’s body) is permitted to do so.

  • body: The InlineFunctionExpr’s FunctionBody.

  • captured context: the static context is the static context of the inline function expression, with the exception of the static context value type which is absentDM40. The dynamic context has an absent focus, and a set of variable bindings comprising the variable values component of the dynamic context of the InlineFunctionExpr.

The following are examples of some inline function expressions:

  • This example creates a function that takes no arguments and returns a sequence of the first 6 primes:

    function() as xs:integer+ { 2, 3, 5, 7, 11, 13 }
  • This example creates a function that takes two xs:double arguments and returns their product:

    fn($a as xs:double, $b as xs:double) as xs:double { $a * $b }
  • This example creates and invokes a function that captures the value of a local variable in its scope:

    let $incrementors := (
      for $x in 1 to 10
      return function($y) as xs:integer { $x + $y }
    )
    return $incrementors[2](4)

    The result of this expression is 6

4.5.2.6 Focus Functions

[Definition: A focus function is an inline function expression in which the function signature is implicit: the function takes a single argument of type item()* (that is, any value), and binds this to the context value when evaluating the function body, which returns a result of type item()*.]

Here are some examples of focus functions:

  • fn { @age } - a function that expects a node as its argument, and returns the @age attribute of that node.

  • fn { . + 1 } - a function that expects a number as its argument, and returns that number plus one.

  • function { `${ . }` } - a function that expects a string as its argument, and prepends a "$" character.

  • function { head(.) + foot(.) } - a function that expects a sequence of numbers as its argument, and returns the sum of the first and last items in the sequence.

Focus functions are often useful as arguments to simple higher-order functions such as fn:sort. For example, to sort employees by salary, write sort(//employee, (), fn { +@salary }). (The unary plus has the effect of converting the attribute’s value to a number, for numeric sorting).

Focus functions can also be useful on the right-hand side of the sequence arrow operator and mapping arrow operator. For example, $s => tokenize() =!> fn { `"{.}"` }() first tokenizes the string $s, then wraps each token in double quotation marks.

The result of calling the function { EXPR } (or fn { EXPR }), with a single argument whose value is $Z arguments, is obtained by evaluating EXPR with a dynamic context in which the context value is $Z, the context position is 1 (one), and the context size is 1 (one).

For example, the expression every(1 to 10, fn { . gt 0 }) returns true.

4.5.2.7 Function Identity

It is sometimes useful to be able to establish whether two variables refer to the same function or to different functions. For this purpose, every function item has an identity. Functions with the same identity are indistinguishable in every way; in particular, any function call with identical arguments will produce an identical result.

In general, evaluation of an expression that returns a function item other than one that was present in its operands delivers a function item whose identity is unique, and thus distinct from any other function item. There are two exceptions to this rule:

  • Evaluating a function reference such as count#1 returns the same function every time. Specifically, if the function name identifies a function definition that is not context dependent (which is the most usual case), then all function references using this function name and arity return the same function. For more details see 4.5.2.4 Named Function References.

  • An optimizer is permitted to rewrite deterministicFO40 expressions in such a way that repeated evaluation is avoided, and this may be done without consideration of function identity. For example, if the expression contains(?, "e") appears within the body of a for clause, or if the same expression is written repeatedly in a query, then an optimizer may decide to evaluate it once only, and thus return the same function item each time. Similarly, if the expression fn($x) { $x + 1 } appears more than once, or is evaluated repeatedly, then it may return the same function each time.

  • In addition, optimizers are allowed to replace any expression with an equivalent expression. For example, count(?) may be rewritten as count#1. Similarly, fn($x) { $x + 1 } may be rewritten as fn($y) { $y + 1 }. This may lead to different expressions returning identical function items.

4.6 Path Expressions

PathExpr ::= ("/" RelativePathExpr?)
| ("//" RelativePathExpr)
| RelativePathExpr
/* xgc: leading-lone-slash */
RelativePathExpr ::= StepExpr (("/" | "//") StepExpr)*

[Definition: A path expression consists of a series of one or more steps, separated by / or //, and optionally beginning with / or //. A path expression is typically used to locate nodes within trees. ]

Absolute path expressions (those starting with an initial / or //), start their selection from the root node of a tree; relative path expressions (those without a leading / or //) start from the context value.

A path expression consisting of a single step is evaluated as described in 4.6.4 Steps.

4.6.1 Absolute Path Expressions

A path expression consisting of / on its own is treated as an abbreviation for /..

An expression of the form /PP (that is, a path expression with a leading /) is treated as an abbreviation for the expression self::node()/(fn:root(.) treat as document-node())/PP. The effect of this expansion is that for every item J in the context value V:

  1. A type error occurs if J is not a node [err:XPTY0020].

  2. The root node R of the tree containing J is selected.

  3. A dynamic error occurs if R is not a document node [err:XPDY0050].

  4. The expression that follows the leading / is evaluated with R as the context value.

The results of these multiple evaluations are then combined into a single sequence; if the result is a set of nodes, the nodes are delivered in document order with duplicates eliminated.

An expression of the form //PP (that is, a path expression with a leading //) is treated as an abbreviation for the expression self::node()/(fn:root(.) treat as document-node())/descendant-or-self:node()/PP. The effect of this expansion is that for every item J in the context value V:

  1. A type error occurs if J is not a node [err:XPTY0020].

  2. The root node R of the tree containing J is selected.

  3. A dynamic error occurs if R is not a document node [err:XPDY0050].

  4. The descendants of R are selected, along with R itself.

  5. For every node D in this set of nodes, the expression that follows the leading // is evaluated with D as the context value.

The results of these multiple evaluations are then combined into a single sequence; if the result is a set of nodes, the nodes are delivered in document order with duplicates eliminated.

If the context value is not a node, a type error is raised [err:XPTY0020]. At evaluation time, if the root node of the context node is not a document node, a dynamic error is raised [err:XPDY0050].

Note:

The descendants of a node do not include attribute nodes or namespace nodes.

Note:

A // on its own is not allowed by the grammar.

4.6.2 Relative Path Expressions

RelativePathExpr ::= StepExpr (("/" | "//") StepExpr)*

A relative path expression is a path expression that selects nodes within a tree by following a series of steps starting at the nodes in the context value (which may be any kind of node, not necessarily the root of the tree).

Each non-initial occurrence of // in a path expression is expanded as described in 4.6.7 Abbreviated Syntax, leaving a sequence of steps separated by /. This sequence of steps is then evaluated from left to right. So a path such as E1/E2/E3/E4 is evaluated as ((E1/E2)/E3)/E4. The semantics of a path expression are thus defined by the semantics of the binary / operator, which is defined in 4.6.3 Path operator (/).

Note:

Although the semantics describe the evaluation of a path with more than two steps as proceeding from left to right, the / operator is in most cases associative, so evaluation from right to left usually delivers the same result. The cases where / is not associative arise when the functions fn:position() and fn:last() are used: A/B/position() delivers a sequence of integers from 1 to the size of (A/B), whereas A/(B/position()) restarts the counting at each B element.

The following example illustrates the use of relative path expressions. In each case it is assumed that the context value is a single node, referred to as the context node.

  • child::div1/child::para

    Selects the para element children of the div1 element children of the context node; that is, the para element grandchildren of the context node that have div1 parents.

Note:

Since each step in a path provides context nodes for the following step, in effect, only the last step in a path is allowed to return a sequence of non-nodes.

Note:

The / character can be used either as a complete path expression or as the beginning of a longer path expression such as /*. Also, * is both the multiply operator and a wildcard in path expressions. This can cause parsing difficulties when / appears on the left-hand side of *. This is resolved using the leading-lone-slash constraint. For example, /* and / * are valid path expressions containing wildcards, but /*5 and / * 5 raise syntax errors. Parentheses must be used when / is used on the left-hand side of an operator that could be confused with a node test, as in (/) * 5. Similarly, 4 + / * 5 raises a syntax error, but 4 + (/) * 5 is a valid expression. The expression 4 + / is also valid, because / does not occur on the left-hand side of the operator.

Similarly, in the expression / union /*, union is interpreted as an element name rather than an operator. For it to be parsed as an operator, the expression should be written (/) union /*.

4.6.3 Path operator (/)

The path operator / is primarily used for locating nodes within trees. Its left-hand operand must return a sequence of nodes. The result of the operator is either a sequence of nodes (in document order, with no duplicates), or a sequence of non-nodes.

The operation E1/E2 is evaluated as follows: Expression E1 is evaluated, and if the result is not a (possibly empty) sequence S of nodes, a type error is raised [err:XPTY0019]. Each node in S then serves in turn to provide an inner focus (the node as the context value, its position in S as the context position, the length of S as the context size) for an evaluation of E2, as described in 2.2.2 Dynamic Context. The sequences resulting from all the evaluations of E2 are combined as follows:

  1. If every evaluation of E2 returns a (possibly empty) sequence of nodes, these sequences are combined, and duplicate nodes are eliminated based on node identity. The resulting node sequence is returned in document order.

  2. If every evaluation of E2 returns a (possibly empty) sequence of non-nodes, these sequences are concatenated, in order, and returned. The returned sequence preserves the orderings within and among the subsequences generated by the evaluations of E2.

  3. If the multiple evaluations of E2 return at least one node and at least one non-node, a type error is raised [err:XPTY0018].

Note:

The semantics of the path operator can also be defined using the simple map operator (!) as follows (the function fn:distinct-ordered-nodes($R) has the effect of eliminating duplicates and sorting nodes into document order):

let $R := E1 ! E2
return if (every $r in $R satisfies $r instance of node())
       then (fn:distinct-ordered-nodes($R))
       else if (every $r in $R satisfies not($r instance of node()))
       then $R
       else error()

For a table comparing the step operator to the map operator, see 4.18 Simple map operator (!).

4.6.4 Steps

StepExpr ::= PostfixExpr | AxisStep
AxisStep ::= (ReverseStep | ForwardStep) Predicate*
ForwardStep ::= (ForwardAxis NodeTest) | AbbrevForwardStep
ReverseStep ::= (ReverseAxis NodeTest) | AbbrevReverseStep
Predicate ::= "[" Expr "]"

[Definition: A step is a part of a path expression that generates a sequence of items and then filters the sequence by zero or more predicates. The value of the step consists of those items that satisfy the predicates, working from left to right. A step may be either an axis step or a postfix expression.] Postfix expressions are described in 4.3 Postfix Expressions.

[Definition: An axis step returns a sequence of nodes that are reachable from a starting node via a specified axis. Such a step has two parts: an axis, which defines the "direction of movement" for the step, and a node test, which selects nodes based on their kind, name, and/or type annotation .]

If the context value is a sequence of zero or more nodes, an axis step returns a sequence of zero or more nodes; otherwise, a type error is raised [err:XPTY0020].

The step expression S is equivalent to ./S. Thus, if the context value is a sequence containing multiple nodes, the semantics of a step expression are equivalent to a path expression in which the step is always applied to a single node. The following description therefore explains the semantics for the case where the context value is a single node, called the context node.

Note:

The equivalence of a step S to the path expression ./S means that the resulting node sequence is returned in document order.

An axis step may be either a forward step or a reverse step, followed by zero or more predicates.

In the abbreviated syntax for a step, the axis can be omitted and other shorthand notations can be used as described in 4.6.7 Abbreviated Syntax.

The unabbreviated syntax for an axis step consists of the axis name and node test separated by a double colon. The result of the step consists of the nodes reachable from the starting node via the specified axis that have the node kind, name, and/or type annotation specified by the node test. For example, the step child::para selects the para element children of the context node: child is the name of the axis, and para is the name of the element nodes to be selected on this axis. The available axes are described in 4.6.4.1 Axes. The available node tests are described in 4.6.4.2 Node Tests. Examples of steps are provided in 4.6.6 Unabbreviated Syntax and 4.6.7 Abbreviated Syntax.

4.6.4.1 Axes

Changes in 4.0  

  1. Four new axes have been defined: preceding-or-self, preceding-sibling-or-self, following-or-self, and following-sibling-or-self.   [Issue 1519 ]

ForwardAxis ::= ("attribute"
| "child"
| "descendant"
| "descendant-or-self"
| "following"
| "following-or-self"
| "following-sibling"
| "following-sibling-or-self"
| "namespace"
| "self") "::"
ReverseAxis ::= ("ancestor"
| "ancestor-or-self"
| "parent"
| "preceding"
| "preceding-or-self"
| "preceding-sibling-or-self") "::"

XPath defines a set of axes for traversing documents, but a host language may define a subset of these axes. The following axes are defined:

  • The child axis contains the children of the context node, which are the nodes returned by the Section 4.3 children AccessorDM40.

    Note:

    Only document nodes and element nodes have children. If the context node is any other kind of node, or if the context node is an empty document or element node, then the child axis is an empty sequence. The children of a document node or element node may be element, processing instruction, comment, or text nodes. Attribute, namespace, and document nodes can never appear as children.

  • The descendant axis is defined as the transitive closure of the child axis; it contains the descendants of the context node (the children, the children of the children, and so on).

    More formally, $node/descendant::node() delivers the result of fn:transitive-closure($node, fn { child::node() }).

  • The descendant-or-self axis contains the context node and the descendants of the context node.

    More formally, $node/descendant-or-self::node() delivers the result of $node/(. | descendant::node()).

  • The parent axis contains the sequence returned by the Section 4.11 parent AccessorDM40, which returns the parent of the context node, or an empty sequence if the context node has no parent.

    Note:

    An attribute node may have an element node as its parent, even though the attribute node is not a child of the element node.

  • The ancestor axis is defined as the transitive closure of the parent axis; it contains the ancestors of the context node (the parent, the parent of the parent, and so on).

    More formally, $node/ancestor::node() delivers the result of fn:transitive-closure($node, fn { parent::node() }).

    Note:

    The ancestor axis includes the root node of the tree in which the context node is found, unless the context node is the root node.

  • The ancestor-or-self axis contains the context node and the ancestors of the context node; thus, the ancestor-or-self axis will always include the root node.

    More formally, $node/ancestor-or-self::node() delivers the result of $node/(. | ancestor::node()).

  • The following-sibling axis contains the context node’s following siblings, that is, those children of the context node’s parent that occur after the context node in document order. If the context node is an attribute or namespace node, the following-sibling axis is empty.

    More formally, $node/following-sibling::node() delivers the result of fn:siblings($node)[. >> $node]).

  • The following-sibling-or-self axis contains the context node, together with the contents of the following-sibling axis.

    More formally, $node/following-sibling-or-self::node() delivers the result of fn:siblings($node)[not(. << $node)]

  • The preceding-sibling axis contains the context node’s preceding siblings, that is, those children of the context node’s parent that occur before the context node in document order. If the context node is an attribute or namespace node, the preceding-sibling axis is empty.

    More formally, $node/preceding-sibling::node() delivers the result of fn:siblings($node)[. << $node].

  • The preceding-sibling-or-self axis contains the context node, together with the contents of the preceding-sibling axis.

    More formally, $node/preceding-sibling-or-self::node() delivers the result of fn:siblings($node)[not(. >> $node).

  • The following axis contains all nodes that are descendants of the root of the tree in which the context node is found, are not descendants of the context node, and occur after the context node in document order.

    More formally, $node/following::node() delivers the result of $node/ancestor-or-self::node()/following-sibling::node()/descendant-or-self::node()

  • The following-or-self axis contains the context node, together with the contents of the following axis.

    More formally, $node/following-or-self::node() delivers the result of $node/(. | following::node()).

  • The preceding axis contains all nodes that are descendants of the root of the tree in which the context node is found, are not ancestors of the context node, and occur before the context node in document order.

    More formally, $node/preceding::node() delivers the result of $node/ancestor-or-self::node()/preceding-sibling::node()/descendant-or-self::node().

  • The preceding-or-self axis contains the context node, together with the contents of the preceding axis.

    More formally, $node/preceding-or-self::node() delivers the result of $node/(. | preceding::node()).

  • The attribute axis contains the attributes of the context node, which are the nodes returned by the Section 4.1 attributes AccessorDM40 ; the axis will be empty unless the context node is an element.

  • The self axis contains just the context node itself.

    The self axis is primarily useful when testing whether the context node satisfies particular conditions, for example if ($x[self::chapter]).

    More formally, $node/self::node() delivers the result of $node.

  • The namespace axis contains the namespace nodes of the context node, which are the nodes returned by the Section 4.7 namespace-nodes AccessorDM40; this axis is empty unless the context node is an element node. The namespace axis is deprecated as of XPath 2.0. If XPath 1.0 compatibility mode is true, the namespace axis must be supported. If XPath 1.0 compatibility mode is false, then support for the namespace axis is implementation-defined. An implementation that does not support the namespace axis when XPath 1.0 compatibility mode is false must raise a static error [err:XPST0010] if it is used. Applications needing information about the in-scope namespaces of an element should use the functions Section 10.2.7 fn:in-scope-prefixesFO40, and Section 10.2.8 fn:namespace-uri-for-prefixFO40.

Axes can be categorized as forward axes and reverse axes. An axis that only ever contains the context node or nodes that are after the context node in document order is a forward axis. An axis that only ever contains the context node or nodes that are before the context node in document order is a reverse axis.

The parent, ancestor, ancestor-or-self, preceding, preceding-or-self, preceding-sibling, and preceding-sibling-or-self axes are reverse axes; all other axes are forward axes.

The ancestor, descendant, following, preceding and self axes partition a document (ignoring attribute and namespace nodes): they do not overlap and together they contain all the nodes in the document.

[Definition: Every axis has a principal node kind. If an axis can contain elements, then the principal node kind is element; otherwise, it is the kind of nodes that the axis can contain.] Thus:

  • For the attribute axis, the principal node kind is attribute.

  • For the namespace axis, the principal node kind is namespace.

  • For all other axes, the principal node kind is element.

4.6.4.2 Node Tests

Changes in 4.0  

  1. If the default namespace for elements and types has the special value ##any, then an unprefixed name in a NameTest acts as a wildcard, matching names in any namespace or none.   [Issue 296 PR 1181 30 April 2024]

[Definition: A node test is a condition on the name, kind (element, attribute, text, document, comment, or processing instruction), and/or type annotation of a node. A node test determines which nodes contained by an axis are selected by a step.]

NodeTest ::= UnionNodeTest | SimpleNodeTest
UnionNodeTest ::= "(" SimpleNodeTest ("|" SimpleNodeTest)* ")"
SimpleNodeTest ::= KindTest | NameTest
NameTest ::= EQName | Wildcard
Wildcard ::= "*"
| (NCName ":*")
| ("*:" NCName)
| (BracedURILiteral "*")
/* ws: explicit */
EQName ::= QName | URIQualifiedName
KindTest ::= DocumentTest
| ElementTest
| AttributeTest
| SchemaElementTest
| SchemaAttributeTest
| PITest
| CommentTest
| TextTest
| NamespaceNodeTest
| AnyKindTest

A UnionNodeTest matches a node N if at least one of the constituent SimpleNodeTests matches N.

For example, (div1|div2|div3) matches a node named div1, div2, or div3

[Definition: A node test that consists only of an EQName or a Wildcard is called a name test.] A name test that consists of an EQName matches a node N if and only if the kind of node N is the principal node kind for the step axis and the expanded QName of the node is equal (as defined by the eq operator) to the expanded QName specified by the name test. For example, child::para selects the para element children of the context node; if the context node has no para children, it selects an empty set of nodes. attribute::abc:href selects the attribute of the context node with the QName abc:href; if the context node has no such attribute, it selects an empty set of nodes.

If the EQName is a lexical QName, it is resolved into an expanded QName using the statically known namespaces in the expression context. It is a static error [err:XPST0081] if the QName has a prefix that does not correspond to any statically known namespace. An unprefixed QName, when used as a name test on an axis whose principal node kind is element, is interpreted as follows:

A name test is not satisfied by an element node whose name does not match the expanded QName of the name test, even if it is in a substitution group whose head is the named element.

A node test * is true for any node of the principal node kind of the step axis. For example, child::* will select all element children of the context node, and attribute::* will select all attributes of the context node.

A node test can have the form NCName:*. In this case, the prefix is expanded in the same way as with a lexical QName, using the statically known namespaces in the static context. If the prefix is not found in the statically known namespaces, a static error is raised [err:XPST0081]. The node test is true for any node of the principal node kind of the step axis whose expanded QName has the namespace URI to which the prefix is bound, regardless of the local part of the name.

A node test can contain a BracedURILiteral, for example Q{http://example.com/msg}*. Such a node test is true for any node of the principal node kind of the step axis whose expanded QName has the namespace URI specified in the BracedURILiteral, regardless of the local part of the name.

A node test can also have the form *:NCName. In this case, the node test is true for any node of the principal node kind of the step axis whose local name matches the given NCName, regardless of its namespace or lack of a namespace.

[Definition: An alternative form of a node test called a kind test can select nodes based on their kind, name, and type annotation.] The syntax and semantics of a kind test are described in 3.1 Sequence Types and 3.1.2 Sequence Type Matching. When a kind test is used in a node test, only those nodes on the designated axis that match the kind test are selected. Shown below are several examples of kind tests that might be used in path expressions:

  • node() matches any node.

  • text() matches any text node.

  • comment() matches any comment node.

  • namespace-node() matches any namespace node.

  • element() matches any element node.

  • schema-element(person) matches any element node whose name is person (or is in the substitution group headed by person), and whose type annotation is the same as (or is derived from) the declared type of the person element in the in-scope element declarations.

  • element(person) matches any element node whose name is person, regardless of its type annotation.

  • element(doctor|nurse) matches any element node whose name is doctor or nurse, regardless of its type annotation.

  • element(person, surgeon) matches any non-nilled element node whose name is person, and whose type annotation is surgeon or is derived from surgeon.

  • element(doctor|nurse, medical-staff) matches any non-nilled element node whose name is doctor or nurse, and whose type annotation is medical-staff or is derived from medical-staff.

  • element(*, surgeon) matches any non-nilled element node whose type annotation is surgeon (or is derived from surgeon), regardless of its name.

  • attribute() matches any attribute node.

  • attribute(price) matches any attribute whose name is price, regardless of its type annotation.

  • attribute(*, xs:decimal) matches any attribute whose type annotation is xs:decimal (or is derived from xs:decimal), regardless of its name.

  • document-node() matches any document node.

  • document-node(element(book)) matches any document node whose children consist of a single element node that satisfies the kind test element(book), interleaved with zero or more comments and processing instructions, and no text nodes.

  • document-node(book) is an abbreviation for document-node(element(book)).

4.6.4.3 Implausible Axis Steps

Changes in 4.0  

  1. The rules for reporting type errors during static analysis have been changed so that a processor has more freedom to report errors in respect of constructs that are evidently wrong, such as @price/@value, even though dynamic evaluation is defined to return an empty sequence rather than an error.   [Issue 602 PR 603 25 July 2023]

Certain axis steps, given an inferred type for the context value, are classified as implausible. During the static analysis phase, a processor may (subject to the rules in 2.4.6 Implausible Expressions) report a static error when such axis steps are encountered: [err:XPTY0144].

More specifically, an axis step is classified as implausible if any of the following conditions applies:

  1. The inferred item type of the context value is a node kind for which the specified axis is always empty: for example, the inferred item type of the context value is attribute() and the axis is child.

  2. The node test exclusively selects node kinds that cannot appear on the specified axis: for example, the axis is child and the node test is document-node().

  3. In a schema-aware environment, when using the child, descendant, descendant-or-self, or attribute axes, the inferred item type of the context value has a content type that does not allow any node matching the node test to be present on the relevant axis. For example, if the inferred item type of the context value is schema-element(list) and the relevant element declaration (taking into account substitution group membership and wildcards) only allows item children, the axis step child::li will never select anything and is therefore classified as implausible.

Examples of implausible axis steps include the following:

  • @code/text(): attributes cannot have text node children.

  • /@code: document nodes cannot have attributes.

  • ancestor::text(): the ancestor axis never returns text nodes.

Note:

Processors may choose not to classify the expression /.. as implausible, since XSLT 1.0 users were sometimes advised to use this construct as an explicit way of denoting the empty sequence.

4.6.5 Predicates within Steps

AxisStep ::= (ReverseStep | ForwardStep) Predicate*
Predicate ::= "[" Expr "]"

A predicate within a AxisStep has similar syntax and semantics to a predicate within a filter expression. The only difference is in the way the context position is set for evaluation of the predicate.

Note:

The operator [] binds more tightly than /. This means that the expression a/b[1] is interpreted as child::a/(child::b[1]): it selects the first b child of every a element, in contrast to (a/b)[1] which selects the first b element that is a child of some a element.

A common mistake is to write //a[1] where (//a)[1] is intended. The first expression, //a[1], selects every descendant a element that is the first child of its parent (it expands to /descendant-or-self::node()/child::a[1]), whereas (//a)[1] selects the a element in the document.

For the purpose of evaluating the context position within a predicate, the input sequence is considered to be sorted as follows: into document order if the predicate is in a forward-axis step, into reverse document order if the predicate is in a reverse-axis step, or in its original order if the predicate is not in a step.

More formally:

  • For a step using a forwards axis, such as child::test[P], the result is the same as for the equivalent filter expression (child::test)[P] (note the parentheses). The same applies if there are multiple predicates, for example child::test[P1][P2][P3] is equivalent to (child::test)[P1][P2][P3].

  • For a step using a reverse axis, such as ancestor::test[P], the result is the same as the expression reverse(ancestor::test)[P] => reverse(). The same applies if there are multiple predicates, for example ancestor::test[P1][P2][P3] is equivalent to reverse(ancestor::test)[P1][P2][P3] => reverse().

Note:

The result of the expression preceding-sibling::* is in document order, but preceding-sibling::*[1] selects the last preceding sibling element, that is, the one that immediately precedes the context node.

Similarly, the expression preceding-sibling::x[1,2,3] selects the last three preceding siblings, returning them in document order. For example, given the input:

<doc><a/><b/><c/><d/><e/><f/></doc>

The result of //e ! preceding-sibling::*[1,2,3] is <b/>, <c/>, <d/>. The expression //e ! preceding-sibling::*[3,2,1] delivers exactly the same result.

Here are some examples of axis steps that contain predicates:

  • This example selects the second chapter element that is a child of the context node:

    child::chapter[2]
  • This example selects all the descendants of the context node that are elements named "toy" and whose color attribute has the value "red":

    descendant::toy[attribute::color = "red"]
  • This example selects all the employee children of the context node that have both a secretary child element and an assistant child element:

    child::employee[secretary][assistant]
  • This example selects the innermost div ancestor of the context node:

    ancestor::div[1]
  • This example selects the outermost div ancestor of the context node:

    ancestor::div[last()]
  • This example selects the names of all the ancestor elements of the context node that have an @id attribute, outermost element first:

    ancestor::*[@id]

Note:

The expression ancestor::div[1] parses as an AxisStep with a reverse axis, and the position 1 therefore refers to the first ancestor div in reverse document order, that is, the innermost div. By contrast, (ancestor::div)[1] parses as a FilterExpr, and therefore returns the first qualifying div element in the order of the ancestor::div expression, which is in document order.

The fact that a reverse-axis step assigns context positions in reverse document order for the purpose of evaluating predicates does not alter the fact that the final result of the step is always in document order.

The expression ancestor::(div1|div2)[1] does not have the same meaning as (ancestor::div1|ancestor::div2)[1]. In the first expression, the predicate [1] is within a step that uses a reverse axis, so nodes are counted in reverse document order. In the second expression, the predicate applies to the result of a union expression, so nodes are counted in document order.

When the context value for evaluation of a step includes multiple nodes, the step is evaluated separately for each of those nodes, and the results are combined without reordering. This means, for example, that if the context value contains three list nodes, and each of those nodes has multiple item children, then the step item[1] will deliver a sequence of three item elements, namely the first item from each list, retaining the order of the respective list elements.

4.6.6 Unabbreviated Syntax

This section provides a number of examples of path expressions in which the axis is explicitly specified in each step. The syntax used in these examples is called the unabbreviated syntax. In many common cases, it is possible to write path expressions more concisely using an abbreviated syntax, as explained in 4.6.7 Abbreviated Syntax.

These examples assume that the context value is a single node, referred to as the context node.

  • child::para selects the para element children of the context node.

  • child::(para|bullet) selects the para and bullet element children of the context node.

  • child::* selects all element children of the context node.

  • child::text() selects all text node children of the context node.

  • child::(text()|comment()) selects all text node and comment node children of the context node.

  • child::node() selects all the children of the context node. Note that no attribute nodes are returned, because attributes are not children.

  • attribute::name selects the name attribute of the context node.

  • attribute::* selects all the attributes of the context node.

  • parent::node() selects the parent of the context node. If the context node is an attribute node, this expression returns the element node (if any) to which the attribute node is attached.

  • descendant::para selects the para element descendants of the context node.

  • ancestor::div selects all div ancestors of the context node.

  • ancestor-or-self::div selects the div ancestors of the context node and, if the context node is a div element, the context node as well.

  • descendant-or-self::para selects the para element descendants of the context node and, if the context node is a para element, the context node as well.

  • self::para selects the context node if it is a para element, and otherwise returns an empty sequence.

  • self::(chapter|appendix) selects the context node if it is a chapter or appendix element, and otherwise returns an empty sequence.

  • child::chapter/descendant::para selects the para element descendants of the chapter element children of the context node.

  • child::*/child::para selects all para grandchildren of the context node.

  • / selects the root of the tree that contains the context node, but raises a dynamic error if this root is not a document node.

  • /descendant::para selects all the para elements in the same document as the context node.

  • /descendant::list/child::member selects all the member elements that have a list parent and that are in the same document as the context node.

  • child::para[position() = 1] selects the first para child of the context node.

  • child::para[position() = last()] selects the last para child of the context node.

  • child::para[position() = last()-1] selects the last but one para child of the context node.

  • child::para[position() > 1] selects all the para children of the context node other than the first para child of the context node.

  • following-sibling::chapter[position() = 1] selects the next chapter sibling of the context node.

  • following-sibling::(chapter|appendix)[position() = 1] selects the next sibling of the context node that is either a chapter or an appendix.

  • preceding-sibling::chapter[position() = 1] selects the previous chapter sibling of the context node.

  • /descendant::figure[position() = 42] selects the forty-second figure element in the document containing the context node.

  • /child::book/child::chapter[position() = 5]/child::section[position() = 2] selects the second section of the fifth chapter of the book whose parent is the document node that contains the context node.

  • child::para[attribute::type eq "warning"] selects all para children of the context node that have a type attribute with value warning.

  • child::para[attribute::type eq 'warning'][position() = 5] selects the fifth para child of the context node that has a type attribute with value warning.

  • child::para[position() = 5][attribute::type eq "warning"] selects the fifth para child of the context node if that child has a type attribute with value warning.

  • child::chapter[child::title = 'Introduction'] selects the chapter children of the context node that have one or more title children whose typed value is equal to the string Introduction.

  • child::chapter[child::title] selects the chapter children of the context node that have one or more title children.

  • child::*[self::chapter or self::appendix] selects the chapter and appendix children of the context node.

  • child::*[self::(chapter|appendix)][position() = last()] selects the last chapter or appendix child of the context node.

4.6.7 Abbreviated Syntax

AbbrevForwardStep ::= ("@" NodeTest) | SimpleNodeTest
AbbrevReverseStep ::= ".."

The abbreviated syntax permits the following abbreviations:

  1. The attribute axis attribute:: can be abbreviated by @. For example, a path expression para[@type = "warning"] is short for child::para[attribute::type = "warning"] and so selects para children with a type attribute with value equal to warning.

  2. If the axis name is omitted from an axis step, the default axis is child, with two exceptions: (1) if the NodeTest in an axis step contains an AttributeTest or SchemaAttributeTest then the default axis is attribute; (2) if the NodeTest in an axis step is a NamespaceNodeTest then the default axis is namespace, but in an implementation that does not support the namespace axis, an error is raised [err:XQST0134].

    Note:

    The namespace axis is deprecated as of XPath 2.0, but is required in some languages that use XPath, including XSLT.

    For example, the path expression section/para is an abbreviation for child::section/child::para, and the path expression section/@id is an abbreviation for child::section/attribute::id. Similarly, section/attribute(id) is an abbreviation for child::section/attribute::attribute(id). Note that the latter expression contains both an axis specification and a node test.

    Note:

    An abbreviated axis step that omits the axis name must use a SimpleNodeTest rather than a UnionNodeTest. This means that a construct such as (ul|ol) is treated as an abbreviation for (child::ul|child::ol) rather than child::(ul|ol). Since the two constructs have exactly the same semantics, this is not actually a restriction.

  3. Each non-initial occurrence of // is effectively replaced by /descendant-or-self::node()/ during processing of a path expression. For example, div1//para is short for child::div1/descendant-or-self::node()/child::para and so will select all para descendants of div1 children.

    Note:

    The path expression //para[1] does not mean the same as the path expression /descendant::para[1]. The latter selects the first descendant para element; the former selects all descendant para elements that are the first para children of their respective parents.

  4. A step consisting of .. is short for parent::node(). For example, ../title is short for parent::node()/child::title and so will select the title children of the parent of the context node.

    Note:

    The expression ., known as a context value reference, is a primary expression, and is described in 4.2.3 Context Value References.

Here are some examples of path expressions that use the abbreviated syntax. These examples assume that the context value is a single node, referred to as the context node:

  • para selects the para element children of the context node.

  • * selects all element children of the context node.

  • text() selects all text node children of the context node.

  • @name selects the name attribute of the context node.

  • @(id|name) selects the id and name attributes of the context node.

  • @* selects all the attributes of the context node.

  • para[1] selects the first para child of the context node.

  • para[last()] selects the last para child of the context node.

  • */para selects all para grandchildren of the context node.

  • /book/chapter[5]/section[2] selects the second section of the fifth chapter of the book whose parent is the document node that contains the context node.

  • chapter//para selects the para element descendants of the chapter element children of the context node.

  • //para selects all the para descendants of the root document node and thus selects all para elements in the same document as the context node.

  • //@version selects all the version attribute nodes that are in the same document as the context node.

  • //list/member selects all the member elements in the same document as the context node that have a list parent.

  • .//para selects the para element descendants of the context node.

  • .. selects the parent of the context node.

  • ../@lang selects the lang attribute of the parent of the context node.

  • para[@type = "warning"] selects all para children of the context node that have a type attribute with value warning.

  • para[@type = "warning"][5] selects the fifth para child of the context node that has a type attribute with value warning.

  • para[5][@type = "warning"] selects the fifth para child of the context node if that child has a type attribute with value warning.

  • chapter[title = "Introduction"] selects the chapter children of the context node that have one or more title children whose typed value is equal to the string Introduction.

  • chapter[title] selects the chapter children of the context node that have one or more title children.

  • employee[@secretary and @assistant] selects all the employee children of the context node that have both a secretary attribute and an assistant attribute.

  • book/(chapter|appendix)/section selects every section element that has a parent that is either a chapter or an appendix element, that in turn is a child of a book element that is a child of the context node.

  • If E is any expression that returns a sequence of nodes, then the expression E/. returns the same nodes in document order, with duplicates eliminated based on node identity.

4.7 Sequence Expressions

XPath 4.0 supports operators to construct, filter, and combine sequences of items. Sequences are never nested—for example, combining the values 1, (2, 3), and ( ) into a single sequence results in the sequence (1, 2, 3).

4.7.1 Sequence Concatenation

Expr ::= (ExprSingle ++ ",")

[Definition: One way to construct a sequence is by using the comma operator, which evaluates each of its operands and concatenates the resulting sequences, in order, into a single result sequence.] Empty parentheses can be used to denote an empty sequence.

A sequence may contain duplicate items, but a sequence is never an item in another sequence. When a new sequence is created by concatenating two or more input sequences, the new sequence contains all the items of the input sequences and its length is the sum of the lengths of the input sequences.

[Definition: The sequence concatenation of a number of sequences S1, S2, ... Sn is defined to be the sequence formed from the items of S1, followed by the items from S2, and so on, retaining order.] The comma operator returns the sequence concatenation of its two operands; repeated application (for example $s1, $s2, $s3, $s4) delivers the sequence concatenation of multiple sequences.

Note:

In places where the grammar calls for ExprSingle, such as the arguments of a function call, any expression that contains a top-level comma operator must be enclosed in parentheses.

Here are some examples of expressions that construct sequences:

  • The result of this expression is a sequence of five integers:

    (10, 1, 2, 3, 4)
  • This expression combines four sequences of length one, two, zero, and two, respectively, into a single sequence of length five. The result of this expression is the sequence 10, 1, 2, 3, 4.

    (10, (1, 2), (), (3, 4))
  • The result of this expression is a sequence containing all salary children of the context node followed by all bonus children.

    (salary, bonus)
  • Assuming that $price is bound to the value 10.50, the result of this expression is the sequence 10.50, 10.50.

    ($price, $price)

4.7.2 Range Expressions

RangeExpr ::= AdditiveExpr ( "to" AdditiveExpr )?

A RangeExpression can be used to construct a sequence of integers. Each of the operands is converted as though it was an argument of a function with the expected parameter type xs:integer?. If either operand is an empty sequence, or if the integer derived from the first operand is greater than the integer derived from the second operand, the result of the range expression is an empty sequence. If the two operands convert to the same integer, the result of the range expression is that integer. Otherwise, the result is a sequence containing the two integer operands and every integer between the two operands, in increasing order.

The following examples illustrate the semantics:

  • 1 to 4 returns the sequence 1, 2, 3, 4

  • 10 to 10 returns the singleton sequence 10

  • 10 to 1 returns the empty sequence

  • -13 to -10 returns the sequence -13, -12, -11, -10

More formally, a RangeExpression is evaluated as follows:

  1. Each of the operands of the to operator is converted as though it was an argument of a function with the expected parameter type xs:integer?.

  2. If either operand is an empty sequence, or if the integer derived from the first operand is greater than the integer derived from the second operand, the result of the range expression is an empty sequence.

  3. If the two operands convert to the same integer, the result of the range expression is that integer.

  4. Otherwise, the result is a sequence containing the two integer operands and every integer between the two operands, in increasing order.

The following examples illustrate the use of RangeExpressions .

This example uses a range expression as one operand in constructing a sequence. It evaluates to the sequence 10, 1, 2, 3, 4.

(10, 1 to 4)

This example selects the first four items from an input sequence:

$input[position() = 1 to 4]

This example returns the sequence (0, 0.1, 0.2, 0.3, 0.5):

$x = (1 to 5)!.*0.1

This example constructs a sequence of length one containing the single integer 10.

10 to 10

The result of this example is a sequence of length zero.

15 to 10

This example uses the fn:reverse function to construct a sequence of six integers in decreasing order. It evaluates to the sequence 15, 14, 13, 12, 11, 10.

reverse(10 to 15)

Note:

To construct a sequence of integers based on steps other than 1, use the fn:slice function, as defined in Section 14.1 General functions and operators on sequences FO31.

4.7.3 Combining Node Sequences

UnionExpr ::= IntersectExceptExpr ( ("union" | "|") IntersectExceptExpr )*
IntersectExceptExpr ::= InstanceofExpr ( ("intersect" | "except") InstanceofExpr )*

XPath 4.0 provides the following operators for combining sequences of nodes:

  • The union and | operators are equivalent. They take two node sequences as operands and return a sequence containing all the nodes that occur in either of the operands.

  • The intersect operator takes two node sequences as operands and returns a sequence containing all the nodes that occur in both operands.

  • The except operator takes two node sequences as operands and returns a sequence containing all the nodes that occur in the first operand but not in the second operand.

All these operators eliminate duplicate nodes from their result sequences based on node identity. The resulting sequence is returned in document order.

If an operand of union, intersect, or except contains an item that is not a node, a type error is raised [err:XPTY0004].

If an IntersectExceptExpr contains more than two InstanceofExprs, they are grouped from left to right. With a UnionExpr, it makes no difference how operands are grouped, the results are the same.

Here are some examples of expressions that combine sequences. Assume the existence of three element nodes that we will refer to by symbolic names A, B, and C. Assume that the variables $seq1, $seq2 and $seq3 are bound to the following sequences of these nodes:

  • $seq1 is bound to (A, B)

  • $seq2 is bound to (A, B)

  • $seq3 is bound to (B, C)

Then:

  • $seq1 union $seq2 evaluates to the sequence (A, B).

  • $seq2 union $seq3 evaluates to the sequence (A, B, C).

  • $seq1 intersect $seq2 evaluates to the sequence (A, B).

  • $seq2 intersect $seq3 evaluates to the sequence containing B only.

  • $seq1 except $seq2 evaluates to the empty sequence.

  • $seq2 except $seq3 evaluates to the sequence containing A only.

In addition to the sequence operators described here, see Section 13 Functions and operators on sequencesFO40 for functions defined on sequences.

4.8 Arithmetic Expressions

Changes in 4.0  

  1. The symbols × and ÷ can be used for multiplication and division.

XPath 4.0 provides arithmetic operators for addition, subtraction, multiplication, division, and modulus, in their usual binary and unary forms.

AdditiveExpr ::= MultiplicativeExpr ( ("+" | "-") MultiplicativeExpr )*
MultiplicativeExpr ::= UnionExpr ( ("*" | "×" | "div" | "÷" | "idiv" | "mod") UnionExpr )*
UnaryExpr ::= ("-" | "+")* ValueExpr
ValueExpr ::= SimpleMapExpr

A subtraction operator must be preceded by whitespace if it could otherwise be interpreted as part of the previous token. For example, a-b will be interpreted as a name, but a - b and a -b will be interpreted as arithmetic expressions. (See A.3.5 Whitespace Rules for further details on whitespace handling.)

The arithmetic operator symbols * and × (xD7) are interchangeable, and denote multiplication.

The arithmetic operator symbols div and ÷ (xF7) are interchangeable, and denote division.

If an AdditiveExpr contains more than two MultiplicativeExprs, they are grouped from left to right. So, for instance,

A - B + C - D

is equivalent to

((A - B) + C) - D

Similarly, the operands of a MultiplicativeExpr are grouped from left to right.

The first step in evaluating an arithmetic expression is to evaluate its operands. The order in which the operands are evaluated is implementation-dependent.

If XPath 1.0 compatibility mode is true, each operand is evaluated by applying the following steps, in order:

  1. Atomization is applied to the operand. The result of this operation is called the atomized operand.

  2. If the atomized operand is an empty sequence, the result of the arithmetic expression is the xs:double value NaN, and the implementation need not evaluate the other operand or apply the operator. However, an implementation may choose to evaluate the other operand in order to determine whether it raises an error.

  3. If the atomized operand is a sequence of length greater than one, any items after the first item in the sequence are discarded.

  4. If the atomized operand is now an instance of type xs:boolean, xs:string, xs:decimal (including xs:integer), xs:float, or xs:untypedAtomic, then it is converted to the type xs:double by applying the fn:number function. (Note that fn:number returns the value NaN if its operand cannot be converted to a number.)

If XPath 1.0 compatibility mode is false, each operand is evaluated by applying the following steps, in order:

  1. Atomization is applied to the operand. The result of this operation is called the atomized operand.

  2. If the atomized operand is an empty sequence, the result of the arithmetic expression is an empty sequence, and the implementation need not evaluate the other operand or apply the operator. However, an implementation may choose to evaluate the other operand in order to determine whether it raises an error.

  3. If the atomized operand is a sequence of length greater than one, a type error is raised [err:XPTY0004].

  4. If the atomized operand is of type xs:untypedAtomic, it is cast to xs:double. If the cast fails, a dynamic error is raised. [err:FORG0001]FO40

After evaluation of the operands, if the types of the operands are a valid combination for the given arithmetic operator, the operator is applied to the operands, resulting in an atomic item or a dynamic error (for example, an error might result from dividing by zero). The combinations of atomic types that are accepted by the various arithmetic operators, and their respective result types, are listed in B.2 Operator Mapping together with the operator functions that define the semantics of the operator for each type combination, including the dynamic errors that can be raised by the operator. The definitions of the operator functions are found in [XQuery and XPath Functions and Operators 4.0].

If the types of the operands, after evaluation, are not a valid combination for the given operator, according to the rules in B.2 Operator Mapping, a type error is raised [err:XPTY0004].

XPath 4.0 provides three division operators:

Here are some examples of arithmetic expressions:

  • The first expression below returns the xs:decimal value -1.5, and the second expression returns the xs:integer value -1:

    -3 div 2
    -3 idiv 2
  • Subtraction of two date values results in a value of type xs:dayTimeDuration:

    $emp/hiredate - $emp/birthdate
  • This example illustrates the difference between a subtraction operator and a hyphen:

    $unit-price - $unit-discount
  • Unary operators have higher precedence than binary operators (other than !, /, and []), subject of course to the use of parentheses. Therefore, the following two examples have different meanings:

    -$bellcost + $whistlecost
    -($bellcost + $whistlecost)

Note:

Multiple consecutive unary arithmetic operators are permitted.

4.9 String Expressions

This section describes several ways of constructing strings.

4.9.1 String Concatenation Expressions

StringConcatExpr ::= RangeExpr ( "||" RangeExpr )*

String concatenation expressions allow the string representations of values to be concatenated. In XPath 4.0, $a || $b is equivalent to fn:concat($a, $b). The following expression evaluates to the string concatenate:

() || "con" || ("cat", "enate")

4.9.2 String Templates

Changes in 4.0  

  1. String templates provide a new way of constructing strings: for example `{$greeting}, {$planet}!` is equivalent to $greeting || ', ' || $planet || '!'   [Issue 58 PR 324 29 January 2023]

StringTemplate ::= "`" (StringTemplateFixedPart | StringTemplateVariablePart)* "`"
/* ws: explicit */
StringTemplateFixedPart ::= ((Char - ('{' | '}' | '`')) | "{{" | "}}" | "``")*
/* ws: explicit */
StringTemplateVariablePart ::= EnclosedExpr
/* ws: explicit */
EnclosedExpr ::= "{" Expr? "}"

String templates provide an alternative way of constructing strings. For example, the expression `Pi is { round(math:pi(), 4) }` returns the string "Pi is 3.1416".

A string template starts and ends with U+0060 (GRAVE ACCENT, BACKTICK, `) , popularly known as a back-tick. Between the back-ticks is a string consisting of an sequence of fixed parts and variable parts:

  • A variable part consists of an optional XPath expression enclosed in curly brackets ({}): more specifically, a string conforming to the XPath production Expr?.

    Note:

    An expression within a variable part may contain an unescaped U+007B (LEFT CURLY BRACKET, {) or U+007D (RIGHT CURLY BRACKET, }) within a StringLiteral or within a comment.

    The fact that the expression is optional means that the string contained between the curly brackets may be zero-length, may comprise whitespace only, or may contain XPath comments. The effective value in this case is a zero-length string, which is equivalent to omitting the variable part entirely, together with its curly-bracket delimiters.

  • A fixed part may contain any characters, except that:

    • The character U+007B (LEFT CURLY BRACKET, {) must be written as {{.

    • The character U+007D (RIGHT CURLY BRACKET, }) must be written as }}.

    • The character U+0060 (GRAVE ACCENT, BACKTICK, `) must be written as ``.

    Following the principles of the “longest token” rule, any occurrence of {{ within the fixed part is interpreted as an escaped left curly bracket. This means that the enclosed expression must not start with U+007B (LEFT CURLY BRACKET, {) : if this is required, the two left curly brackets can be separated by whitespace. For example the string template `{{"key":"{ {1:"yes", 0:"no"}?$condition}"}}` evaluates to the string {"key":"yes"} or {"key":"no"} depending on the value of $condition.

    By contrast, if the enclosed expression ends with U+007D (RIGHT CURLY BRACKET, }) , this can be immediately followed by the closing U+007D (RIGHT CURLY BRACKET, }) delimiter without intervening whitespace.

The result of evaluating a string template is the string obtained by concatenating the expansions of the fixed and variable parts:

  • The expansion of a fixed part is obtained by replacing any double curly brackets ({{ or }}) by the corresponding single curly bracket, and replacing doubled back-ticks (``) by a single back-tick.

  • The expansion of a variable part containing an expression is as follows:

    1. Atomization is applied to the value of the enclosed expression, converting it to a sequence of atomic items.

    2. If the result of atomization is an empty sequence, the result is the zero-length string. Otherwise, each atomic item in the atomized sequence is cast into a string.

    3. The individual strings resulting from the previous step are merged into a single string by concatenating them with a single space character between each pair.

  • The expansion of an empty variable part (one that contains no expression) is a zero-length string.

For example:

let $greeting := "Hello",
    $planet := "Mars"
return `{ $greeting }, { $planet }!`

returns "Hello, Mars!".

The expression:

let $longMonths := (1, 3, 5, 7, 8, 10, 12)
return `The months with 31 days are: { $longMonths }.`

returns "The months with 31 days are: 1 3 5 7 8 10 12.".

Note:

The rules for processing an enclosed expression are identical to the rules for attributes in XQuery direct element constructors. These rules differ slightly from the rules in XSLT attribute value templates, where adjacent text nodes are concatenated with no separator, prior to atomization.

Note:

A string template containing no variable parts is effectively just another way of writing a string literal: "Goethe", 'Goethe', and `Goethe` are interchangeable. This means that back-ticks can sometimes be a useful way of delimiting a string that contains both single and double quotes: `He said: "I didn't."`.

It is sometimes useful to use string templates in conjunction with the fn:char function to build strings containing special characters, for example `Chapter{ fn:char("nbsp") }{ $chapNr }`.

Note:

String literals containing an ampersand behave differently between XPath and XQuery: in XPath (unless first expanded by an XML parser) the string literal "Bacon & Eggs" represents a string containing an ampersand, while in XQuery it is an error, because an ampersand is taken as introducing a character reference. This difference does not arise for string templates, since neither XPath nor XQuery recognizes entity or character references in a string template. This means that back-tick delimited strings (such as `Bacon & Eggs`) may be useful in contexts where an XPath expression is required to have the same effect whether it is evaluated using an XPath or an XQuery processor.

In XQuery, the token ``[ is recognized as the start of a StringConstructor, under the “longest token” rule (see A.3 Lexical structure). This means that the construct ``[1] is not recognized as a StringTemplate followed by a predicate. Although the token ``[ is not used in XPath, it is reserved for compatibility reasons, and must be rejected as syntactically invalid. In the unlikely event that an empty StringTemplate followed by a predicate is wanted, whitespace or parentheses can be used to avoid the tokenization problem.

4.10 Comparison Expressions

Comparison expressions allow two values to be compared. XPath 4.0 provides three kinds of comparison expressions, called value comparisons, general comparisons, and node comparisons.

ComparisonExpr ::= OtherwiseExpr ( (ValueComp
| GeneralComp
| NodeComp) OtherwiseExpr )?
ValueComp ::= "eq" | "ne" | "lt" | "le" | "gt" | "ge"
GeneralComp ::= "=" | "!=" | "<" | "<=" | ">" | ">="
NodeComp ::= "is" | "<<" | ">>"

Note:

When an XPath expression is written within an XML document, the XML escaping rules for special characters must be followed; thus < must be written as &lt;.

For a summary of the differences between different ways of comparing atomic items in XPath 4.0, see H Atomic Comparisons: An Overview.

4.10.1 Value Comparisons

Changes in 4.0  

  1. The rules for value comparisons when comparing values of different types (for example, decimal and double) have changed to be transitive. A decimal value is no longer converted to double, instead the double is converted to a decimal without loss of precision. This may affect compatibility in edge cases involving comparison of values that are numerically very close.

The value comparison operators are eq, ne, lt, le, gt, and ge. Value comparisons are used for comparing single values.

The first step in evaluating a value comparison is to evaluate its operands. The order in which the operands are evaluated is implementation-dependent. Each operand is evaluated by applying the following steps, in order:

  1. Atomization is applied to each operand. The result of this operation is called the atomized operand.

  2. If an atomized operand is an empty sequence, the result of the value comparison is an empty sequence, and the implementation need not evaluate the other operand or apply the operator. However, an implementation may choose to evaluate the other operand in order to determine whether it raises an error.

  3. If an atomized operand is a sequence of length greater than one, a type error is raised [err:XPTY0004].

  4. If an atomized operand is of type xs:untypedAtomic, it is cast to xs:string.

    Note:

    The purpose of this rule is to make value comparisons transitive. Users should be aware that the general comparison operators have a different rule for casting of xs:untypedAtomic operands. Users should also be aware that transitivity of value comparisons may be compromised by loss of precision during type conversion (for example, two xs:integer values that differ slightly may both be considered equal to the same xs:float value because xs:float has less precision than xs:integer).

  5. If the two operands are instances of different primitive types (meaning the 19 primitive types defined in Section 3.2 Primitive datatypesXS2), then:

    1. If each operand is an instance of one of the types xs:string or xs:anyURI, then both operands are cast to type xs:string.

    2. If each operand is an instance of one of the types xs:decimal or xs:float, then both operands are cast to type xs:float.

    3. If each operand is an instance of one of the types xs:decimal, xs:float, or xs:double, then both operands are cast to type xs:double.

    4. Otherwise, a type error is raised [err:XPTY0004].

      Note:

      The primitive type of an xs:integer value for this purpose is xs:decimal.

  6. Finally, if the types of the operands are a valid combination for the given operator, the operator is applied to the operands.

The combinations of atomic types that are accepted by the various value comparison operators, and their respective result types, are listed in B.2 Operator Mapping together with the operator functions that define the semantics of the operator for each type combination. The definitions of the operator functions are found in [XQuery and XPath Functions and Operators 4.0].

Informally, if both atomized operands consist of exactly one atomic item, then the result of the comparison is true if the value of the first operand is (equal, not equal, less than, less than or equal, greater than, greater than or equal) to the value of the second operand; otherwise the result of the comparison is false.

If the types of the operands, after evaluation, are not a valid combination for the given operator, according to the rules in B.2 Operator Mapping, a type error is raised [err:XPTY0004].

Here are some examples of value comparisons:

  • The following comparison atomizes the node(s) that are returned by the expression $book/author. The comparison is true only if the result of atomization is the value "Kennedy" as an instance of xs:string or xs:untypedAtomic. If the result of atomization is an empty sequence, the result of the comparison is an empty sequence. If the result of atomization is a sequence containing more than one value, a type error is raised [err:XPTY0004].

    $book1/author eq "Kennedy"
  • The following comparison is true because atomization converts an array to its member sequence:

    [ "Kennedy" ] eq "Kennedy"
  • The following path expression contains a predicate that selects products whose weight is greater than 100. For any product that does not have a weight subelement, the value of the predicate is the empty sequence, and the product is not selected. This example assumes that weight is a validated element with a numeric type.

    //product[weight gt 100]
  • The following comparison is true if my:hatsize and my:shoesize are both user-defined types that are derived by restriction from a primitive numeric type:

    my:hatsize(5) eq my:shoesize(5)
  • The following comparison is true. The eq operator compares two QNames by performing codepoint-comparisons of their namespace URIs and their local names, ignoring their namespace prefixes.

    QName("http://example.com/ns1", "this:color") eq
    QName("http://example.com/ns1", "that:color")

4.10.2 General Comparisons

Changes in 4.0  

  1. Operators such as < and > can use the full-width forms and to avoid the need for XML escaping.

The general comparison operators are =, !=, <, <=, >, and >=. General comparisons are existentially quantified comparisons that may be applied to operand sequences of any length. The result of a general comparison that does not raise an error is always true or false.

If XPath 1.0 compatibility mode is true, a general comparison is evaluated by applying the following rules, in order:

  1. If either operand is a single atomic item that is an instance of xs:boolean, then the other operand is converted to xs:boolean by taking its effective boolean value.

  2. Atomization is applied to each operand. After atomization, each operand is a sequence of atomic items.

  3. If the comparison operator is <, <=, >, or >=, then each item in both of the operand sequences is converted to the type xs:double by applying the fn:number function. (Note that fn:number returns the value NaN if its operand cannot be converted to a number.)

  4. The result of the comparison is true if and only if there is a pair of atomic items, one in the first operand sequence and the other in the second operand sequence, that have the required magnitude relationship. Otherwise the result of the comparison is false or an error. The magnitude relationship between two atomic items is determined by applying the following rules. If a cast operation called for by these rules is not successful, a dynamic error is raised. [err:FORG0001]FO40

    1. If at least one of the two atomic items is an instance of a numeric type, then both atomic items are converted to the type xs:double by applying the fn:number function.

    2. If at least one of the two atomic items is an instance of xs:string, or if both atomic items are instances of xs:untypedAtomic, then both atomic items are cast to the type xs:string.

    3. If one of the atomic items is an instance of xs:untypedAtomic and the other is not an instance of xs:string, xs:untypedAtomic, or any numeric type, then the xs:untypedAtomic item is cast to the dynamic type of the other value.

    4. After performing the conversions described above, the atomic items are compared using one of the value comparison operators eq, ne, lt, le, gt, or ge, depending on whether the general comparison operator was =, !=, <, <=, >, or >=. The values have the required magnitude relationship if and only if the result of this value comparison is true.

If XPath 1.0 compatibility mode is false, a general comparison is evaluated by applying the following rules, in order:

  1. Atomization is applied to each operand. After atomization, each operand is a sequence of atomic items.

  2. The result of the comparison is true if and only if there is a pair of atomic items, one in the first operand sequence and the other in the second operand sequence, that have the required magnitude relationship. Otherwise the result of the comparison is false or an error. The magnitude relationship between two atomic items is determined by applying the following rules. If a cast operation called for by these rules is not successful, a dynamic error is raised. [err:FORG0001]FO40

    1. If both atomic items are instances of xs:untypedAtomic, then the values are cast to the type xs:string.

    2. If exactly one of the atomic items is an instance of xs:untypedAtomic, it is cast to a type depending on the other value’s dynamic type T according to the following rules, in which V denotes the value to be cast:

      1. If T is a numeric type or is derived from a numeric type, then V is cast to xs:double.

      2. If T is xs:dayTimeDuration or is derived from xs:dayTimeDuration, then V is cast to xs:dayTimeDuration.

      3. If T is xs:yearMonthDuration or is derived from xs:yearMonthDuration, then V is cast to xs:yearMonthDuration.

      4. In all other cases, V is cast to the primitive base type of T.

      Note:

      The special treatment of the duration types is required to avoid errors that may arise when comparing the primitive type xs:duration with any duration type.

    3. After performing the conversions described above, the atomic items are compared using one of the value comparison operators eq, ne, lt, le, gt, or ge, depending on whether the general comparison operator was =, !=, <, <=, >, or >=. The values have the required magnitude relationship if and only if the result of this value comparison is true.

When evaluating a general comparison in which either operand is a sequence of items, an implementation may return true as soon as it finds an item in the first operand and an item in the second operand that have the required magnitude relationship. Similarly, a general comparison may raise a dynamic error as soon as it encounters an error in evaluating either operand, or in comparing a pair of items from the two operands. As a result of these rules, the result of a general comparison is not deterministic in the presence of errors.

Here are some examples of general comparisons:

  • The following comparison is true if the typed value of any author subelement of $book1 is "Kennedy" as an instance of xs:string or xs:untypedAtomic:

    $book1/author = "Kennedy"
  • The following comparison is true because atomization converts an array to its member sequence:

    [ "Obama", "Nixon", "Kennedy" ] = "Kennedy"
  • The following example contains three general comparisons. The value of the first two comparisons is true, and the value of the third comparison is false. This example illustrates the fact that general comparisons are not transitive.

    (1, 2) = (2, 3)
    (2, 3) = (3, 4)
    (1, 2) = (3, 4)
  • The following example contains two general comparisons, both of which are true. This example illustrates the fact that the = and != operators are not inverses of each other.

    (1, 2) = (2, 3)
    (1, 2) != (2, 3)
  • Suppose that $a, $b, and $c are bound to element nodes with type annotation xs:untypedAtomic, with string values "1", "2", and "2.0" respectively. Then ($a, $b) = ($c, 3.0) returns false, because $b and $c are compared as strings. However, ($a, $b) = ($c, 2.0) returns true, because $b and 2.0 are compared as numbers.

4.10.3 Node Comparisons

Node comparisons are used to compare two nodes, by their identity or by their document order. The result of a node comparison is defined by the following rules:

  1. The operands of a node comparison are evaluated in implementation-dependent order.

  2. If either operand is an empty sequence, the result of the comparison is an empty sequence, and the implementation need not evaluate the other operand or apply the operator. However, an implementation may choose to evaluate the other operand in order to determine whether it raises an error.

  3. Each operand must be either a single node or an empty sequence; otherwise a type error is raised [err:XPTY0004].

  4. A comparison with the is operator is true if the two operand nodes are the same node; otherwise it is false. See [XQuery and XPath Data Model (XDM) 4.0] for the definition of node identity.

  5. A comparison with the << operator returns true if the left operand node precedes the right operand node in document order; otherwise it returns false.

  6. A comparison with the >> operator returns true if the left operand node follows the right operand node in document order; otherwise it returns false.

Here are some examples of node comparisons:

  • The following comparison is true only if the left and right sides each evaluate to exactly the same single node:

    /books/book[isbn = "1558604820"] is /books/book[call = "QA76.9 C3845"]
  • The following comparison is true only if the node identified by the left side occurs before the node identified by the right side in document order:

    /transactions/purchase[parcel = "28-451"] << /transactions/sale[parcel = "33-870"]

4.11 Logical Expressions

A logical expression is either an and-expression or an or-expression. If a logical expression does not raise an error, its value is always one of the boolean values true or false.

OrExpr ::= AndExpr ( "or" AndExpr )*
AndExpr ::= ComparisonExpr ( "and" ComparisonExpr )*

The first step in evaluating a logical expression is to find the effective boolean value of each of its operands (see 2.5.4 Effective Boolean Value).

The value of an and-expression is determined by the effective boolean values (EBVs) of its operands, as shown in the following table:

AND: EBV2 = true EBV2 = false error in EBV2
EBV1 = true true false error
EBV1 = false false false if XPath 1.0 compatibility mode is true, then false; otherwise either false or error.
error in EBV1 error if XPath 1.0 compatibility mode is true, then error; otherwise either false or error. error

The value of an or-expression is determined by the effective boolean values (EBVs) of its operands, as shown in the following table:

OR: EBV2 = true EBV2 = false error in EBV2
EBV1 = true true true if XPath 1.0 compatibility mode is true, then true; otherwise either true or error.
EBV1 = false true false error
error in EBV1 if XPath 1.0 compatibility mode is true, then error; otherwise either true or error. error error

If XPath 1.0 compatibility mode is true, the order in which the operands of a logical expression are evaluated is effectively prescribed. Specifically, it is defined that when there is no need to evaluate the second operand in order to determine the result, then no error can occur as a result of evaluating the second operand.

If XPath 1.0 compatibility mode is false, the order in which the operands of a logical expression are evaluated is implementation-dependent. In this case, an or-expression can return true if the first expression evaluated is true, and it can raise an error if evaluation of the first expression raises an error. Similarly, an and-expression can return false if the first expression evaluated is false, and it can raise an error if evaluation of the first expression raises an error. As a result of these rules, a logical expression is not deterministic in the presence of errors, as illustrated in the examples below.

Here are some examples of logical expressions:

In addition to and- and or-expressions, XPath 4.0 provides a function named fn:not that takes a general sequence as parameter and returns a boolean value. The fn:not function is defined in [XQuery and XPath Functions and Operators 4.0]. The fn:not function reduces its parameter to an effective boolean value. It then returns true if the effective boolean value of its parameter is false, and false if the effective boolean value of its parameter is true. If an error is encountered in finding the effective boolean value of its operand, fn:not raises the same error.

4.12 For and Let Expressions

XPath provides two closely-related expressions, called For and Let expressions, that can be used to bind variables to values. The complete syntax is shown here, and relevant parts of the syntax are repeated in subsequent sections of this document.

ForExpr ::= ForClause ForLetReturn
LetExpr ::= LetClause ForLetReturn
ForClause ::= "for" (ForBinding ++ ",")
ForBinding ::= ForItemBinding | ForMemberBinding | ForEntryBinding
ForItemBinding ::= VarNameAndType PositionalVar? "in" ExprSingle
ForMemberBinding ::= "member" VarNameAndType PositionalVar? "in" ExprSingle
ForEntryBinding ::= ((ForEntryKeyBinding ForEntryValueBinding?) | ForEntryValueBinding) PositionalVar? "in" ExprSingle
ForEntryKeyBinding ::= "key" VarNameAndType
ForEntryValueBinding ::= "value" VarNameAndType
LetClause ::= "let" (LetBinding ++ ",")
LetBinding ::= VarNameAndType ":=" ExprSingle
VarNameAndType ::= "$" EQName TypeDeclaration?
TypeDeclaration ::= "as" SequenceType
PositionalVar ::= "at" VarName
ForLetReturn ::= ForExpr | LetExpr | ("return" ExprSingle)

4.12.1 For Expressions

Changes in 4.0  

  1. A for member clause is added to FLWOR expressions to allow iteration over an array.  [Issue 49 PR 344 10 February 2023]

  2. Multiple for and let clauses can be combined in an expression without an intervening return keyword.   [Issue 22 PR 28 18 December 2020]

  3. A for key/value clause is added to FLWOR expressions to allow iteration over maps.  [Issue 31 PR 1249 1 June 2024]

  4. A positional variable can be defined in a for expression.   [Issue 231 PR 1131 1 April 2024]

  5. The type of a variable used in a for expression can be declared.   [Issue 796 PR 1131 1 April 2024]

XPath provides an iteration facility called a for expression. It can be used to iterate over the items of a sequence, the members of an array, or the entries in a map.

ForExpr ::= ForClause ForLetReturn
ForClause ::= "for" (ForBinding ++ ",")
ForBinding ::= ForItemBinding | ForMemberBinding | ForEntryBinding
ForItemBinding ::= VarNameAndType PositionalVar? "in" ExprSingle
ForMemberBinding ::= "member" VarNameAndType PositionalVar? "in" ExprSingle
ForEntryBinding ::= ((ForEntryKeyBinding ForEntryValueBinding?) | ForEntryValueBinding) PositionalVar? "in" ExprSingle
ForLetReturn ::= ForExpr | LetExpr | ("return" ExprSingle)
TypeDeclaration ::= "as" SequenceType
PositionalVar ::= "at" VarName

A for expression is evaluated as follows:

  1. If the ForClause includes multiple ForBindings with a comma separator, the forexpression is first expanded to a set of nested for expressions, each of which contains a single ForBinding. More specifically, every separating comma is replaced by for.

    For example, the expression for $x in X, $y in Y return $x + $y is expanded to for $x in X for $y in Y return $x + $y.

  2. Having performed this expansion, variables bound in the ForClause are called the range variables, the variable named in the PositionalVar (if present) is called the position variable, the expression that follows the in keyword is called the binding expression, and the expression in the ForLetReturn part (that is, the following LetExpr or ForExpr, or the ExprSingle that follows the return keyword) is called the return expression.

    [Definition: The result of evaluating the binding expression in a for expression is called the binding collection ].

  3. If a position variable is declared, its type is implicitly xs:integer. Its name (as a QName) must be different from the name of a range variable declared in the same ForBinding. [err:XQST0089].

  4. When a ForItemBinding is used (that is, when none of the keywords member, key, or value is used), the expression iterates over the items in a sequence:

    1. If a TypeDeclaration is present then each item in the binding collection is converted to the specified type by applying the coercion rules.

    2. The return expression is evaluated once for each item in the binding collection, with a dynamic context in which the range variable is bound to that item, and the position variable (if present) is bound to the one-based position of that item in the binding collection, as an instance of type xs:integer.

    3. The result of the for expression is the sequence concatenation of the results of the successive evaluations of the return expression.

  5. When the member keyword is present:

    1. The value of the binding collection must be a single array. Otherwise, a type error is raised: [err:XPTY0141].

    2. If a TypeDeclaration is present then each member of the binding collection array is converted to the specified type by applying the coercion rules. (Recall that this can be any sequence, not necessarily a single item).

    3. The result of the single-variable for member expression is obtained by evaluating the return expression once for each member of that array, with the range variable bound to that member

    4. The return expression is evaluated once for each member of the binding collection array, with a dynamic context in which the range variable is bound to that member, and the position variable (if present) is bound to the one-based position of that member in the binding collection.

    5. The result of the for expression is the sequence concatenation of the results of the successive evaluations of the return expression.

      Note that the result is a sequence, not an array.

  6. When the key and/or value keywords are present:

    1. The value of the binding collection must be a single map. Otherwise, a type error is raised: [err:XPTY0141]. The map is treated as a sequence of key/value pairs, in implementation dependent order.

    2. If the key keyword is present, then the corresponding variable is bound to the key part of the key/value pair.

    3. If the value keyword is present, then the corresponding variable is bound to the value part of the key/value pair.

    4. If both the key and value keywords are present, then the corresponding variables must have distinct names. [err:XQST0089].

    5. If a TypeDeclaration is present for the key, then each key is converted to the specified type by applying the coercion rules.

    6. If a TypeDeclaration is present for the value, then each value is converted to the specified type by applying the coercion rules.

    7. The result of the single-variable for key/value expression is obtained by evaluating the return expression once for each entry in the map, with the range variables bound to that entry as described.

    8. The return expression is evaluated once for each entry of the binding collection map, with a dynamic context in which the key range variable (if present) is bound to the key part of that entry, the value range variable (if present) is bound to the value part of that entry, and the position variable (if present) is bound to the one-based position of that entry in the implementation dependent ordering of the binding collection.

    9. The result of the for expression is the sequence concatenation of the results of the successive evaluations of the return expression.

      Note that the result is a sequence, not a map.

The following example illustrates the use of a for expression in restructuring an input document. The example is based on the following input:

<bib>
  <book>
    <title>TCP/IP Illustrated</title>
    <author>Stevens</author>
    <publisher>Addison-Wesley</publisher>
  </book>
  <book>
    <title>Advanced Programming in the Unix Environment</title>
    <author>Stevens</author>
    <publisher>Addison-Wesley</publisher>
  </book>
  <book>
    <title>Data on the Web</title>
    <author>Abiteboul</author>
    <author>Buneman</author>
    <author>Suciu</author>
  </book>
</bib>

The following example transforms the input document into a list in which each author’s name appears only once, followed by a list of titles of books written by that author. This example assumes that the context value is the bib element in the input document.

for $a in distinct-values(book/author)
return ((book/author[. = $a])[1], book[author = $a]/title)

The result of the above expression consists of the following sequence of elements. The titles of books written by a given author are listed after the name of the author. The ordering of author elements in the result is implementation-dependent due to the semantics of the fn:distinct-values function.

<author>Stevens</author>
<title>TCP/IP Illustrated</title>
<title>Advanced Programming in the Unix environment</title>
<author>Abiteboul</author>
<title>Data on the Web</title>
<author>Buneman</author>
<title>Data on the Web</title>
<author>Suciu</author>
<title>Data on the Web</title>

The following example illustrates a for expression containing more than one variable:

for $i in (10, 20),
    $j in (1, 2)
return ($i + $j)

The result of the above expression, expressed as a sequence of numbers, is as follows: 11, 12, 21, 22

The scope of a variable bound in a for expression is the return expression. The scope does not include the expression to which the variable is bound. The following example illustrates how a variable binding may reference another variable bound earlier in the same for expression:

for $x in $z, $y in f($x)
return g($x, $y)

The following example illustrates processing of an array.

for member $map in parse-json('[{ "x": 1, "y": 2 }, { "x": 10, "y": 20 }]') 
return $map ! (?x + ?y)

The result is the sequence (3, 30).

The following example illustrates processing of a map.

for key $key value $value in { "x": 1, "y": 2, "z: 3 }
return `{$key}={$value}`

The result is the sequence ("x=1", "y=2", "z=3") (but not necessarily in that order).

Note:

The focus for evaluation of the return clause of a for expression is the same as the focus for evaluation of the for expression itself. The following example, which attempts to find the total value of a set of order-items, is therefore incorrect:

sum(for $i in order-item return @price * @qty)

Instead, the expression must be written to use the variable bound in the for clause:

sum(for $i in order-item return $i!(@price * @qty))

Note:

XPath 4.0 allows the format:

for $order in //orders
for $line in $order/order-line
return $line/value

primarily because it is familiar to XQuery users, some of whom may regard it as more readable than the XPath 3.1 alternative which uses a comma in place of the second for.

4.12.2 Let Expressions

Changes in 4.0  

  1. Multiple for and let clauses can be combined in an expression without an intervening return keyword.   [Issue 22 PR 28 18 December 2020]

  2. The type of a variable used in a let expression can be declared.   [Issue 796 PR 1131 1 April 2024]

XPath allows a variable to be declared and bound to a value using a let expression.

LetExpr ::= LetClause ForLetReturn
LetClause ::= "let" (LetBinding ++ ",")
LetBinding ::= VarNameAndType ":=" ExprSingle
ForLetReturn ::= ForExpr | LetExpr | ("return" ExprSingle)
TypeDeclaration ::= "as" SequenceType

A let expression is evaluated as follows:

  • If the let expression uses multiple variables, it is first expanded to a set of nested let expressions, each of which uses only one variable. Specifically, any separating comma is replaced by let.

    For example, the expression let $x := 4, $y := 3 return $x + $y is expanded to let $x := 4 let $y := 3 return $x + $y.

  • In a single-variable let expression, the variable is called the range variable. The expression that follows the := symbol is evaluated, and if a TypeDeclaration is present, its value is converted to the specified type by applying the coercion rules. The resulting value is called the binding sequence. The expression in the ForLetReturn part (that is, the following LetExpr or ForExpr, or the ExprSingle that follows the return keyword) is called the return expression. The result of the let expression is obtained by evaluating the return expression with a dynamic context in which the range variable is bound to the binding sequence.

The scope of a variable bound in a let expression is the return expression. The scope does not include the expression to which the variable is bound. The following example illustrates how a variable binding may reference another variable bound earlier in the same let expression:

let $x := doc('a.xml')/*, $y := $x//*
return $y[@value gt $x/@min]

Note:

It is not required that the variables should have distinct names. It is permitted, for example, to write:

let $x := "[A fine romance]"
let $x := substring-after($x, "[")
let $x := substring-before($x, "]")
return upper-case($x)

which returns the result "A FINE ROMANCE". Note that this expression declares three separate variables which happen to have the same name; it should not be read as declaring a single variable and binding it successively to different values.

4.13 Maps and Arrays

Most modern programming languages have support for collections of key/value pairs, which may be called maps, dictionaries, associative arrays, hash tables, keyed lists, or objects (these are not the same thing as objects in object-oriented systems). In XPath 4.0, we call these maps. Most modern programming languages also support ordered lists of values, which may be called arrays, vectors, or sequences. In XPath 4.0, we have both sequences and arrays. Unlike sequences, an array is an item, and can appear as an item in a sequence.

In previous versions of the language, element structures and sequences were the only complex data structures. We are adding maps and arrays to XPath 4.0 in order to provide lightweight data structures that are easier to optimize and less complex to use for intermediate processing and to allow programs to easily combine XML processing with JSON processing.

Note:

The XPath 4.0 specification focuses on syntax provided for maps and arrays, especially constructors and lookup.

Some of the functionality typically needed for maps and arrays is provided by functions defined in Section 17 MapsFO40 and Section 18 ArraysFO40, including functions used to read JSON to create maps and arrays, serialize maps and arrays to JSON, combine maps to create a new map, remove map entries to create a new map, iterate over the keys of a map, convert an array to create a sequence, combine arrays to form a new array, and iterate over arrays in various ways.

4.13.1 Maps

[Definition: A map is a function that associates a set of keys with values, resulting in a collection of key / value pairs.] [Definition: Each key / value pair in a map is called an entry.] [Definition: The value associated with a given key is called the associated value of the key.]

4.13.1.1 Map Constructors

Changes in 4.0  

  1. In map constructors, the keyword map is now optional, so map { 0: false(), 1: true() } can now be written { 0: false(), 1: true() }, provided it is used in a context where this creates no ambiguity.   [Issue 1070 PR 1071 26 March 2024]

A map can be created using a MapConstructor.

MapConstructor ::= "map"? "{" (MapConstructorEntry ** ",") "}"
MapConstructorEntry ::= MapKeyExpr ":" MapValueExpr
MapKeyExpr ::= ExprSingle
MapValueExpr ::= ExprSingle

Note:

The keyword map was required in earlier versions of the language; in XPath 4.0 it becomes optional. There may be cases where using the keyword improves readability.

In order to allow the map keyword to be omitted, an incompatible change has been made to XQuery computed element and attribute constructors: if the name of the constructed element or attribute is a language keyword, it must now be written in quotes, for example element "div" {}.

Although the grammar allows a MapConstructor to appear within an EnclosedExpr (that is, between curly brackets), this may be confusing to readers, and using the map keyword in such cases may improve clarity. In any event, if the EnclosedExpr appears in a context such as a StringTemplate, the two adjacent left opening braces must at least be separated by whitespace.

Note:

In some circumstances, it is necessary to include whitespace before or after the colon of a MapConstructorEntry to ensure that it is parsed as intended.

For instance, consider the expression {a:b}. Although it matches the EBNF for MapConstructor (with a matching MapKeyExpr and b matching MapValueExpr), the "longest possible match" rule requires that a:b be parsed as a QName, which results in a syntax error. Changing the expression to {a :b} or {a: b} will prevent this, resulting in the intended parse.

Similarly, consider these three expressions:

{a:b:c}
{a:*:c}
{*:b:c}

In each case, the expression matches the EBNF in two different ways, but the “longest possible match” rule forces the parse in which the MapKeyExpr is a:b, a:*, or *:b (respectively) and the MapValueExpr is c. To achieve the alternative parse (in which the MapKeyExpr is merely a or *), insert whitespace before and/or after the first colon.

See A.3 Lexical structure.

The value of the expression (whether or not the map keyword is present) is a map whose entries correspond to the key-value pairs obtained by evaluating the successive MapKeyExpr and MapValueExpr expressions.

Each MapKeyExpr expression is evaluated and atomized; a type error [err:XPTY0004] occurs if the result is not a single atomic item. The associated value is the result of evaluating the corresponding MapValueExpr. If the MapValueExpr evaluates to a node, the associated value is the node itself, not a new node with the same values. [Definition: Two atomic items K1 and K2 have the same key value if fn:atomic-equal(K1, K2) returns true, as specified in Section 13.2.1 fn:atomic-equalFO40 ] If two or more entries have the same key value then a dynamic error is raised [err:XQDY0137]. The error may be raised statically if two or more entries can be determined statically to have the same key value.

Example: Constructing a fixed map

The following expression constructs a map with seven entries:

{
  "Su" : "Sunday",
  "Mo" : "Monday",
  "Tu" : "Tuesday",
  "We" : "Wednesday",
  "Th" : "Thursday",
  "Fr" : "Friday",
  "Sa" : "Saturday"
}

 

Example: Constructing nested maps

Maps can nest, and can contain any XDM value. Here is an example of a nested map with values that can be string values, numeric values, or arrays:


{
  "book": {
    "title": "Data on the Web",
    "year": 2000,
    "author": [
      {
        "last": "Abiteboul",
        "first": "Serge"
      },
      {
        "last": "Buneman",
        "first": "Peter"
      },
      {
        "last": "Suciu",
        "first": "Dan"
      }
    ],
    "publisher": "Morgan Kaufmann Publishers",
    "price": 39.95
  }
}

Note:

The syntax deliberately mimics JSON, but there are a few differences. JSON constructs that are not accepted in XPath 4.0 map constructors include the keywords true, false, and null, and backslash-escaped characters such as "\n" in string literals. In an XPath 4.0 map constructor, of course, any literal value can be replaced with an expression.

Both the keys and the values in a map constructor can be supplied as expressions rather than as constants; the number of entries in the map, however, is fixed. There are several functions that can be used to construct maps with a variable number of entries:

  • map:build takes any sequence as input, and for each item in the sequence, it computes a key and a value, by calling user-supplied functions.

  • map:merge takes a sequence of maps (often but not necessarily single-entry maps) and merges them into a single map.

  • map:of-pairs takes a sequence of key-value pair mapsFO40 and merges them into a single map.

Any of these functions can be used to build an index of employee elements using the value of the @id attribute as a key:

  • map:build(//employee, fn{@id})

  • map:merge(//employee ! {@id, .})

  • map:of-pairs(//employee ! {'key':@id, 'value':.})

All three functions also provide control over how duplicate keys are handled, allowing them to be used to create a map in which items are grouped according to the value of a grouping key.

4.13.1.2 Maps as Functions

Maps are function items, and a dynamic function call can be used to look up the value associated with a key in a map. If $map is a map and $key is a key, then $map($key) is equivalent to map:get($map, $key). The semantics of such a function call are formally defined in Section 17.3.9 map:getFO40.

Examples:

Note:

XPath 4.0 also provides an alternate syntax for map and array lookup that is more terse, supports wildcards, and allows lookup to iterate over a sequence of maps or arrays. See 4.13.3 Lookup Expressions for details.

Map lookups can be chained.

Examples: (These examples assume that $b is bound to the books map from the previous section)

  • The expression $b("book")("title") returns the string Data on the Web.

  • The expression $b("book")("author") returns the array of authors.

  • The expression $b("book")("author")(1)("last") returns the string Abiteboul.

    (This example combines 4.13.2.2 Arrays as Functions with map lookups.)

4.13.2 Arrays

4.13.2.1 Array Constructors

[Definition: An array is a function item that associates a set of positions, represented as positive integer keys, with values.] The first position in an array is associated with the integer 1. [Definition: The values of an array are called its members.] In the type hierarchy, array has a distinct type, which is derived from function. Atomization converts arrays to sequences (see Atomization).

An array is created using an ArrayConstructor.

ArrayConstructor ::= SquareArrayConstructor | CurlyArrayConstructor
SquareArrayConstructor ::= "[" (ExprSingle ** ",") "]"
CurlyArrayConstructor ::= "array" EnclosedExpr

If a member of an array is a node, its node identity is preserved. In both forms of an ArrayConstructor, if a member expression evaluates to a node, the associated value is the node itself, not a new node with the same values. If the member expression evaluates to a map or array, the associated value is a new map or array with the same values.

A SquareArrayConstructor consists of a comma-delimited set of argument expressions. It returns an array in which each member contains the value of the corresponding argument expression.

Examples:

  • [ 1, 2, 5, 7 ] creates an array with four members: 1, 2, 5, and 7.

  • [ (), (27, 17, 0) ] creates an array with two members: () and the sequence (27, 17, 0).

  • [ $x, local:items(), <tautology>It is what it is.</tautology> ] creates an array with three members: the value of $x, the result of evaluating the function call, and a tautology element.

A CurlyArrayConstructor can use any expression to create its members. It evaluates its operand expression to obtain a sequence of items and creates an array with these items as members. Unlike a SquareArrayConstructor, a comma in a CurlyArrayConstructor is the comma operator, not a delimiter.

Examples:

  • array { $x } creates an array with one member for each item in the sequence to which $x is bound.

  • array { local:items() } creates an array with one member for each item in the sequence to which local:items() evaluates.

  • array { 1, 2, 5, 7 } creates an array with four members: 1, 2, 5, and 7.

  • array { (), (27, 17, 0) } creates an array with three members: 27, 17, and 0.

  • array { $x, local:items(), <tautology>It is what it is.</tautology> } creates an array with the following members: the items to which $x is bound, followed by the items to which local:items() evaluates, followed by a tautology element.

Note:

XPath 4.0 does not provide explicit support for sparse arrays. Use integer-valued maps to represent sparse arrays, for example: { 27 : -1, 153 : 17 }.

4.13.2.2 Arrays as Functions

Arrays are function items, and a dynamic function call can be used to look up the value associated with position in an array. If $array is an array and $index is an integer corresponding to a position in the array, then $array($key) is equivalent to array:get($array, $key). The semantics of such a function call are formally defined in Section 18.2.11 array:getFO40.

Examples:

  • [ 1, 2, 5, 7 ](4) evaluates to 7.

  • [ [ 1, 2, 3 ], [ 4, 5, 6 ] ](2) evaluates to [ 4, 5, 6 ].

  • [ [ 1, 2, 3 ], [ 4, 5, 6 ] ](2)(2) evaluates to 5.

  • [ 'a', 123, <name>Robert Johnson</name> ](3) evaluates to <name>Robert Johnson</name>.

  • array { (), (27, 17, 0) }(1) evaluates to 27.

  • array { (), (27, 17, 0) }(2) evaluates to 17.

  • array { "licorice", "ginger" }(20) raises a dynamic error [err:FOAY0001]FO40.

Note:

XPath 4.0 also provides an alternate syntax for map and array lookup that is more terse, supports wildcards, and allows lookup to iterate over a sequence of maps or arrays. See 4.13.3 Lookup Expressions for details.

4.13.3 Lookup Expressions

Changes in 4.0  

  1. The lookup operator ? can now be followed by a string literal, for cases where map keys are strings other than NCNames. It can also be followed by a variable reference.

  2. A deep lookup operator ?? is provided for searching trees of maps and arrays.   [Issue 297 PR 837 23 November 2023]

  3. Lookup expressions can now take a modifier (such as keys, values, or pairs) enabling them to return structured results rather than a flattened sequence.   [Issues 960 1094 PR 1125 23 April 2024]

XPath 4.0 provides two lookup operators ? and ?? for maps and arrays. These provide a terse syntax for accessing the entries in a map or the members of an array.

The operator "?", known as the shallow lookup operator, returns values found immediately in the operand map or array. The operator "??", known as the deep lookup operator, also searches nested maps and arrays. The effect of the deep lookup operator "??" is explained in 4.13.3.3 Deep Lookup.

4.13.3.1 Postfix Lookup Expressions
LookupExpr ::= PostfixExpr Lookup
Lookup ::= ("?" | "??") (Modifier "::")? KeySpecifier
Modifier ::= "pairs" | "keys" | "values" | "items"
KeySpecifier ::= NCName | IntegerLiteral | StringLiteral | VarRef | ParenthesizedExpr | LookupWildcard
LookupWildcard ::= "*"

A Lookup has two parts: the KeySpecifier determines which entries (in a map) or members (in an array) are selected, and the Modifier determines how they are delivered in the result. The default modifier is items, which delivers the result as a flattened sequence of items.

To take a simple example, given $A as an array [ ("a", "b"), ("c", "d"), ("e", "f") ], some example Lookup expressions are:

Example Lookup Expressions on an Array
Expression Result
$A?* (or $A?items::*)
("a", "b", "c", "d", "e", "f")
$A?pairs::*
({ "key": 1, "value": ("a", "b") }, 
 { "key": 2, "value": ("c", "d") }, 
 { "key": 3, "value": ("e", "f") })
$A?values::*
([ "a", "b" ], [ "c", "d" ], [ "e", "f" ])
$A?keys::*
(1, 2, 3)
$A?2 (or $A?items::2)
("c", "d")
$A?pairs::2
({ "key": 2, "value":("c", "d") })
$A?values::2
([ "c", "d" ])
$A?keys::2
(2)
$A?(3, 1) (or $A?items::(3, 1))
("e", "f", "a", "b")
$A?pairs::(3, 1)
({ "key": 3, "value": ("e", "f") }, 
 { "key": 1, "value": ("a", "b") })
$A?values::(3, 1)
([ "e", "f" ][ "a", "b" ])
$A?keys::(3, 1)
(3, 1)

Similarly, given $M as a map { "X": ("a", "b"), "Y": ("c", "d"), "Z": ("e", "f") }, some example lookup expressions are as follows. Note that because maps are unordered, the results are not necessarily in the order shown.

Example Lookup Expressions on a Map
Expression Result
$M?* (or $M?items::*)
("a", "b", "c", "d", "e", "f")
$M?pairs::*
({ "key": "X", "value": ("a", "b") }, 
 { "key": "Y", "value": ("c", "d") }, 
 { "key": "Z", "value": ("e", "f") })
$M?values::*
([ "a", "b" ], [ "c", "d" ], [ "e", "f" ])
$M?keys::*
("X", "Y", "Z")
$M?Y (or $M?items::Y)
("c", "d")
$M?pairs::Y
({ "key": "Y", "value":("c", "d") })
$M?values::Y
([ "c", "d" ])
$M?keys::Y
("Y")
$M?("Z", "X") (or $A?items::("Z", "X"))
("e", "f", "a", "b")
$M?pairs::("Z", "X")
({ "key": "Z", "value": ("e", "f") }, 
 { "key": "X", "value": ("a", "b") })
$M?values::("Z", "X")
([ "e", "f" ][ "a", "b" ])
$M?keys::("Z", "X")
("Z", "X")

The semantics of a postfix lookup expression E?pairs::KS are defined as follows. The results with other modifiers can be derived from this result, as explained below.

  1. E is evaluated to produce a value $V.

  2. If $V is not a singleton (that is if count($V) ne 1), then the result (by recursive application of these rules) is the value of for $v in $V return $v?pairs::KS.

  3. If $V is a singleton array (that is, if $V instance of array(*)) then:

    1. If KS is a ParenthesizedExpr, then it is evaluated to produce a value $K and the result is:

      data($K) ! { "key": ., "value": array:get($V, .) }

      Note:

      The focus for evaluating the key specifier expression is the same as the focus for the Lookup expression itself.

    2. If the