<?xml version="1.0"?>
<a>
<b/> <!-- this is the //b[count(./*)=0] -->
<b> <!-- this is the //b[./c] -->
<c/>
</b>
<b> <!-- this is the //b[3] -->
<!-- and also //b[./c] -->
<c/>
</b>
</a>
(C) 2019 Masaryk University -- Tomáš Pitner, Luděk Bártek, Adam Rambousek
XPath is a syntax used to specify parts of XML documents (primitive values, nodes, sequences of values or nodes
XPath does not allow to specify parts of text nodes.
Its name is derived from path expression providing a means of hierarchic addressing of the nodes in an XML tree.
XPath uses syntax similar to file system path.
XPath offers standard functions library, as well as user defined functions in either some XPath 2.0 or even XPath 1.x processors.
XPath does not use XML syntax (it would be too long).
XPath 1.0 (revised Sep 7, 2015),
XML Path Language (XPath) 3.0 are W3C Recommendations (Apr 08, 2014)
XPath 3.1 is a Recommendation (Mar 21, 2017)
Backward compatibility: nearly all XPath 1.0 expressions continue to deliver the same result with XPath 3.0 (for exceptions see http://www.w3.org/TR/xpath-30/#id-backwards-compatibility)
XPath is used as a base for XSLT since version 1.0 and
in XQuery since XPath version 2.0.
Zvon XPath 1.0 Tutorial to learn step by step (by Miloslav Nič)
PathEnq XPath 2.0 online evaluator — nice for try&see
Online XPath Tester - Evaluator by ExtendsClass
<?xml version="1.0"?>
<a>
<b/> <!-- this is the //b[count(./*)=0] -->
<b> <!-- this is the //b[./c] -->
<c/>
</b>
<b> <!-- this is the //b[3] -->
<!-- and also //b[./c] -->
<c/>
</b>
</a>
Select the 3rd node b: //b[3]
Select a node "b", which has a child node "c": //b[./c]
Select an empty (eg. no child elements) node b: //b[count(./*)=0]
Select nodes that have to be processed next: <xsl:apply-templates match="para"/>
Select value: <xsl:value-of select="para/@id"/>
(F)or part, eg. for $para in $doc//para
selects all para
in the document doc
(L)et part, eg. let $mypara := $doc//para[@owner='myself']
(W)here part, eg. where $para[@class='task']
(O)rder part, eg. order by $para/@created
Schematron
XML Schema
Path describes (or "navigates" to) an XML document location. Paths syntax is constructed a similar way to paths in file systems, i.e.:
related to a context node (CN), see further, or
related to the root element but predicates are evaluated in relation to CN.
Since XPath 3.0 unified with the XML Schema and XQuery datatypes
XQuery and XPath Data Model 3.0, W3C Recommendation 08 April 2014
Axes (singular axis, plural axes) are sets of document elements, related to (usually relatively) to context.
Context is formed by a document and the current (context) node (CN).
child
contains direct child nodes of CN
descendant
contains all descendants of CN except attributes.
parent
contains the CN parent nod (if it exists)
ancestor
contains all ancestors of CN - means parents, grandparents, etc to a root element (if the CN is not the root element itself)
following-sibling
contains all following siblings of CN (the axis is empty for NS and attributes)
preceding-sibling
dtto, but it contains the preceding sibling.
following
contains all nodes following the CN (except the attributes, child nodes and NS nodes)
preceding
dtto, but contains preceding nodes (except ancestors, attributes, NS)
attribute
contains attributes (for elements only)
namespace
contains all NS nodes of CN (for elements only)
self
the CN itself
descendant-or-self
contains the union of descendant and self axes
ancestor-or-self
contains the union of ancestor and self axes
It is possible to try evaluation of XPath expressions upon a provided XML document by using many online testers without the need of (local PC) installation.
Such as http://codebeautify.org/Xpath-Tester or
PathEnq XPath 2.0 online evaluator
XPath online tester also allows to evaluate XPath against an XML document
//b/child::*
<?xml version="1.0"?>
<a>
<b/>
<b>
<c/> <!-- this "c" will be selected -->
</b>
<b>
<c/> <!-- and this "c" too -->
</b>
</a>
//b/descendant::*
<?xml version="1.0"?>
<a>
<b/>
<b>
<c> <!-- everything "under b" will be selected -->
<d/> <!-- i.e. this "d" too -->
</c>
</b>
<b>
<c/> <!-- and this "c" too -->
</b>
</a>
//d/parent::*
<?xml version="1.0"?>
<a>
<b/>
<b>
<c> <!-- this "c" is the parent of "d" -->
<d/>
</c>
</b>
<b>
<c/>
</b>
</a>
//d/ancestor::*
<?xml version="1.0"?>
<a> <!-- this "a" is ancestor of "d" -->
<b/>
<b> <!-- this "b" is ancestor of "d" -->
<c> <!-- this "c" is ancestor of "d" -->
<d/>
</c>
</b>
<b>
<c/>
</b>
</a>
//b/following-sibling::*
<?xml version="1.0"?>
<a>
<b/> <!-- every child of "a" after this "b" is following-sibling -->
<b> <!-- this "b" too -->
<c>
<d/>
</c>
</b>
<b> <!-- this "b" too -->
<c/>
</b>
</a>
//b/preceding-sibling::*
<?xml version="1.0"?>
<a>
<b/> <!-- this "b" too -->
<b>
<c>
<d/>
</c>
</b> <!-- this "b" is preceding-sibling -->
<b> <!-- every child of "a" before this "b" is preceding-sibling -->
<c/>
</b>
</a>
/a/b/c/following::*
<?xml version="1.0"?>
<a>
<b/>
<b>
<c>
<d/>
</c> <!-- every element starting after "c" is following -->
<e/> <!-- such as this "e" -->
</b>
<b> <!-- this "b" too -->
<c/> <!-- this "c" too -->
</b>
</a>
/a/b/e/preceding::*
<?xml version="1.0"?>
<a>
<b/> <!-- this "b" too -->
<b> <!-- this "b" too -->
<c> <!-- this "c" too -->
<d/> <!-- this "d" too -->
</c>
</b>
/a/b/e/preceding::*
<b>
<d/> <!-- such as this "d" -->
<e/> <!-- every element starting before "e" is preceding -->
</b>
</a>
Figure: /article/para[3]
— selects the 3rd paragraph (element para
) of
article (element article
)
Simplest predicate expression is proximity position specification — see
preceding
.
Attention at reverse axes (ancestor
, preceding
, …) - position is
numbered always from the context node, means opposite to
document physical location directions.
Position specification 3 can be replace by the expression position()=3
.
Used in predicates for calculations. Expressions may contain XPath functions. Expressions may operate on:
text strings
numbers (floating-point numbers)
logical values (boolean)
nodes
sequences.
para
selects all child nodes of context node with name para
*
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
@*
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
/doc/chapter[5]/section[2]
selects the second section
of the fifth chapter
of the doc
chapter//para
selects all descendants of element chapter
with name para
//para
selects all elements para
in the document
//olist/item
selects all elements item
with parent element olist
.//para
selects all descendant nodes of the context node with name para
..
selects the parent node of the context node
../@lang
selects a lang
attribute of the context node parent node
Most common used short notation is at child axis
we use article/para instead of child::article/child::para
.
at attribute:we use para[@type="warning"]
instead of
child::para[attribute::type="warning"]
The next used short notation is //
instead of
/descendant-or-self::node()/
and of course shortcuts .
and ..
For clarity, we keep sometimes the longer form: Do not fight it at all costs!
Final specification available at http://www.w3.org/TR/xpath20/
Different point of view on return values of XPatch expressions: everything is a sequence (even containing a single element) → removes the set node order problems
Introduces conditional expressions and cycles.
Introduces user-defined functions (dynamically evaluate XPath expressions)
Users can uses general and existential quantifiers, for example exist student/name="Fred"
, all student/@id
For more details see http://www.saxonica.com/, pages contains the XPath/XSLT/XQuery processor Saxon as well.
Programming in XPath 3.0 (D. Novatchev)
XPath functions (Mozilla)