XPointer: Fragment Identifiers for XML

XPointer allows links to point to specific parts of an XML document, not just the whole document. It uses XPath expressions to identify the target. Used with XLink to link to fragments of remote XML documents.

Remark: No browser supports XPointer, however, it is used in other XML languages.

Syntax ​

The general form is a URL followed by a # fragment identifier:

url#pointer

The pointer part has 2 forms:

1. Targets an element by its ID directly.

   txt

      dogbreeds.xml#Rottweiler

2. Scheme-based: Wraps an XPath expression in xpointer(...).

   txt

      dogbreeds.xml#xpointer(id('Rottweiler'))
      dogbreeds.xml#xpointer(/dogbreeds/dog[1])

Note that id() is an XPath function that matches attributes declared as type ID in a DTD or xml:id attributes, not just any attribute named "id".

Multiple fragments can be provided: The schemes are evaluated left to right, the first match wins.

dogbreeds.xml#Rottweiler#xpointer(/dogbreeds/dog[1])

The context node in XPointer is always the document root, so relative XPath expressions evaluate from the root and are effectively the same as absolute ones.

Example ​

Target document (dogbreeds.xml):

<dogbreeds>
  <dog breed="Rottweiler" id="Rottweiler">
    <history>The Rottweiler's ancestors were...</history>
  </dog>
  <dog breed="FCRetriever" id="FCRetriever">
    <history>One of the earliest uses of...</history>
  </dog>
</dogbreeds>

Linking document:

<mydogs xmlns:xlink="http://www.w3.org/1999/xlink">
  <mydog>
    <fact xlink:type="simple" xlink:href="dogbreeds.xml#Rottweiler">
      Fact about Rottweiler
    </fact>
  </mydog>
</mydogs>

Remark: XPointer is to XML what #fragment is to HTML, but it is more powerful because it uses XPath, so we can target any node, not just elements with id attributes.