xpath3

Written as xpath3(argnum flags):options in a transform expression.

Minimum arguments	1
Maximum arguments	1
Default arguments	1
Argument modes	SINGLE_PRIMITIVE

Documentation

Extract data from XML/HTML content using XPath3 expressions.

Features:

• Supports standard XPath 3.0/3.1 syntax
• Works with both XML and HTML content
• Automatic recovery (with ‘RECOVER XML’ flag) of broken XML. However, use this flag with caution, as this process may not fully recover the original XML document structure when ambiguilty exists.
• Namespace support:
  • Enabled by default. Can be disabled with ‘IGNORE NAMESPACE’ flag. When ‘IGNORE NAMESPACE’ flag is turned on, namespace mapping is ignored, and the transform’s options data is interpretted as the xpath expression
  • [Optional] Namespace mapping via a preliminary mapping dict before the actual xpath expression, such as {"ns1":"http://mymodel.org/model1/1.0"}||//main//model1

The XPath3 expression (from the transform’s options data) supports 2 forms: - {namespace_mapping}||{xpath_expr}: e.g. {"ns1":"http://mymodel.org/model1/1.0"}||//main//ns1:model1 - {xpath_expr}: e.g. //main//model1

where `{namespace_mapping}` is the namespace mapping dict and `{xpath_expr}` is the actual xpath expression.

Flags

Flag	Letter
spread_output	S
noncaptive_input	I
recover_xml	r
IGNORE_NAMESPACE	N

Examples

Extract all links from HTML

Input args: ["<html><body><a href=\"link1\">Link 1</a><a href=\"link2\">Link 2</a></body></html>"]
Transforms: "xpath3://a/@href"
Output: ["link1","link2"]

Extract text content

Input args: ["<root><item>First</item><item>Second</item></root>"]
Transforms: "xpath3://item/text()"
Output: ["First","Second"]

Extract attribute value

Input args: ["<div id=\"main\" class=\"container\"></div>"]
Transforms: "xpath3://div/@class"
Output: ["container"]

Use xpath3 built-in functions

Input args: ["<root/>"]
Transforms: "xpath3:math:atan(1.0e0)"
Output: 0.7853981633974483

Use xpath3 “let” expression

Input args: ["<root/>"]
Transforms: "xpath3:let $x := 5, $y := 10 return $x + $y"
Output: 15

Use xpath3 conditional expression

Input args: ["<root/>"]
Transforms: "xpath3:if (//item) then 'has items' else 'no items'"
Output: ["no items"]

Use xpath3 sequence construction

Input args: ["<root/>"]
Transforms: "xpath3:(1 to 5)"
Output: [1,2,3,4,5]

Use xpath3 higher order functions

Input args: ["<root/>"]
Transforms: "xpath3:let $f := function($x) { $x * 2 } return $f(5)"
Output: 10

Use xpath3 string-join functions

Input args: ["<root/>"]
Transforms: "xpath3:string-join(('a','b','c'), '-')"
Output: "a-b-c"

Use xpath3 array/map construction

Input args: ["<root/>"]
Transforms: "xpath3:array { 1, 2, 3 }, map { 'a': 1, 'b': 2 }"
Output: [[1,2,3],{"a":1,"b":2}]

Use xpath3 quantified expressions

Input args: ["<root/>"]
Transforms: "xpath3:some $x in (1,2,3) satisfies $x > 2"
Output: true

Use xpath3 arrow operator

Input args: ["<root/>"]
Transforms: "xpath3:('a','b','c') => count()"
Output: 3

Recover slightly malformed xml

Input args: ["<Parent><Child1/><Child2></Parent>"]
Transforms: "xpath3(r):/Parent/*"
Output: ["<Child1/>","<Child2/>"]

Use namespaced xml

Input args: ["<Parent xmlns=\"http://testxml.org/parent/1.0\"><Child1/><Child2/></Parent>"]
Transforms: "xpath3(r):{\"ns1\":\"http://testxml.org/parent/1.0\"}||/ns1:Parent/*"
Output: ["<Child1 xmlns=\"http://testxml.org/parent/1.0\"/>","<Child2 xmlns=\"http://testxml.org/parent/1.0\"/>"]

Ignore namespaces in xml

Input args: ["<Parent xmlns=\"http://testxml.org/parent/1.0\"><Child1/><Child2/></Parent>"]
Transforms: "xpath3(N):/Parent/*"
Output: ["<Child1 xmlns=\"http://testxml.org/parent/1.0\"/>","<Child2 xmlns=\"http://testxml.org/parent/1.0\"/>"]