XSLT and XQuery Serialization 4.0

9.1 Serializing Items

An individual item is serialized as follows:

• A JNode is serialized by serializing its ·content· property.

• An array item in the input tree is serialized to a JSON array by outputting the serialized JSON value of each member within the array separated by delimiters according to the JSON array syntax, i.e. [member, member, ...]. Each member in the array is to be serialized by recursively applying the rules in this section.

• A map item in the input tree is serialized to a JSON object by outputting, for each key/value pair, the string value of the key to a JSON string, followed by the serialized JSON value of the entry, separated by delimiters according to the JSON object syntax, i.e. {key:value, key:value, ...}. The key/value pairs in the serialized output retain the entry order^DM of entries in the map, unless canonical is true, in which case map entries must be sorted according to the rules of [RFC8785].

  Note:

  These rules require sorting according to the UTF-16 representation of the string, which is not the same (in the presence of surrogate pairs) as sorting using the XPath codepoint collation.

  If any two keys of the map item have the same string value, serialization error [err:SERE0022] is raised, unless the allow-duplicate-names parameter is true and the canonical parameter is false.

• A node in the input tree is serialized to a JSON string by outputting the result of serializing the node using the method specified by the json-node-output-method parameter. If the value of the canonical parameter is false, the node is serialized with the serialization parameter omit-xml-declaration set to true and with no other serialization parameters set, otherwise (canonical is true) all parameters are inherited by the sub-serialization process.

• An atomic value^XP of type xs:numeric is serialized to a JSON number.

  If the canonical parameter is true, then the value is cast to type xs:double and the result is serialized according to the rules of [RFC8785] (and by extension [RFC7493]).If the canonical parameter is true, then the value is cast to type xs:double and the result is serialized according to the rules of [RFC8785] (and by extension [RFC7493]). A serialization error occurs for the values NaN, INF, and -INF which cannot be represented in canonical JSON.If the canonical parameter is true, then the value is cast to type xs:double and the result is serialized according to the rules of [RFC8785] (and by extension [RFC7493]). A serialization error occurs for the values NaN, INF, and -INF which cannot be represented in canonical JSON.

  If the canonical parameter is false then:

  • A value of type xs:decimal (including xs:integer) is output in the format that results from casting the value to xs:string.

  • A value of type xs:float is first cast to type xs:double, and then output in the same way as an xs:double.

  • Implementations may serialize an xs:double value using any lexical representation of a JSON number defined in [RFC 7159], but it is recommended to use the same representation as when the canonical parameter is true.Implementations may serialize an xs:double value using any lexical representation of a JSON number defined in [RFC 7159], but it is recommended to use the same representation as when the canonical parameter is true; except that negative zero should be output with a leading minus sign.Implementations may serialize an xs:double value using any lexical representation of a JSON number defined in [RFC 7159], but it is recommended to use the same representation as when the canonical parameter is true; except that negative zero should be output with a leading minus sign.

  • The value NaN is serialized as the JSON token null.

  • The values positive and negative infinity are serialized as 1e9999 and -1e9999 respectively.

    Note:

    These values are permitted by the JSON grammar, but some JSON parsers may reject them. For interoperability, they should be avoided.

• An atomic item^XP of type xs:boolean is serialized to the JSON token true or false.

• An atomic item^XP of type xs:QName in the input tree whose namespace part is "http://www.w3.org/2005/xpath-functions" and whose local part is "null" is serialized to the JSON token null.

  Note:

  This rule is introduced in 4.0, along with an option in the fn:parse-json function to allow a user-defined representation of the JSON value null. While the default representation of null as an empty sequence is usable in many circumstances, an explicit representation of null as a recognizable item can make some operations on JSON-derived values easier.

• Any other atomic value^XP in the input tree is serialized to a JSON string by outputting the result of applying the fn:string function to the item.

• An empty sequence in the input tree is serialized to the JSON token null.

• A sequence of length greater than one in the input tree will result in a serialization error [err:SERE0023].

• Any item in the input tree of a type not specified in the above list will result in a serialization error [err:SERE0021].

[Definition: Whenever a value is serialized to a JSON string, the following procedure is applied to the supplied string:]

1. If the canonical parameter is true, then the string is output as specified in [RFC8785]. The string is first normalized if requested using the normalization-form parameter. The serialization parameters escape-solidus and use-character-maps are ignored.

   Note:

   The escaping rules are restated here for convenience:

   • If the Unicode value falls within the traditional ASCII control character range (U+0000 (NULL) through U+001F (IS1) ), it must be serialized using lowercase hexadecimal Unicode notation (\uhhhh) unless it is in the set of predefined JSON control characters U+0008 (BACKSPACE) , U+0009 (TAB) , U+000A (NEWLINE) , U+000C (FORM FEED) , or U+000D (CARRIAGE RETURN) , which must be serialized as \b, \t, \n, \f, and \r, respectively.

   • If the Unicode value is outside of the ASCII control character range, it must be serialized “as is” unless it is equivalent to U+005C (REVERSE SOLIDUS, BACKSLASH, \) or U+0022 (QUOTATION MARK, ") , which must be serialized as \\ and \", respectively.

   Note that the C1 control characters (codepoints 127-159) are not included in this list.

   Note:

   In canonical JSON, property values within an object are sorted based on their UTF16 representation. This corresponds to the default sort order in languages such as Javascript, Java, and C#, but in the presence of codepoints above 65535, it is not the same as the order produced by Unicode codepoint collation.

2. Otherwise (when canonical is false):

   1. Any character in the string for which character mapping is defined (see 11 Character Maps) is substituted by the replacement string defined in the character map.

   2. Any other character in the input string (but not a character produced by character mapping) is a candidate for Unicode Normalization if requested by the normalization-form parameter, and JSON escaping. JSON escaping replaces the characters U+005C (REVERSE SOLIDUS, BACKSLASH, \) , U+0008 (BACKSPACE) , U+000C (FORM FEED) , U+000A (NEWLINE) , U+000D (CARRIAGE RETURN) , U+0009 (TAB) , or U+005C (REVERSE SOLIDUS, BACKSLASH, \) by the corresponding JSON escape sequences \", \b, \f, \n, \r, \t, or \\ respectively, and any other codepoint in the range 1-31 or 127-159 by an escape in the form \uHHHH where HHHH is the hexadecimal representation of the codepoint value. Escaping further replaces the solidus character (/) by the escape sequence \/ if the escape-solidus parameter is set to true, but not if it is set to false. Escaping is also applied to any characters that cannot be represented in the selected encoding.

3. The resulting string is enclosed in double quotation marks.

Finally, encoding, as controlled by the encoding parameter, converts the character stream produced by the preceding rules into an octet stream. The encoding parameter is ignored if canonical is true.

9.3 The Influence of Serialization Parameters upon the JSON Output Method

When nodes are serialized using the JSON output method, serialization is delegated to the output method specified by the json-node-output-method serialization parameter. The omit-xml-declaration parameter is set to true, and no other serialization parameters are passed down to the serialization method responsible for serializing the node.

The serialization parameters that affect the JSON output method are listed in the following subsections.

Serialization parameters other than those listed are not applicable to this output method. It is the responsibility of the host language to specify whether an error occurs if such a parameter is specified in combination with the JSON output method, or whether the parameter is ignored, or whether it is validated and then ignored.