Regex – Using RegEx in XSLT


I need to parse Visual Studio automatically generated XML documentation to create a report. I decided to use XSLT but I'm very new to it and need help.
Common template is:

    <member name="F:MyNamespace">
      <summary>Some text</summary>

I want to isolate members with name which begins on some word, for example, P:Interfaces.Core. I decided to use RegExp in select statement.

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet version="1.0" xmlns:xsl=""
    <xsl:template match="/" >
        <html xmlns="">
            <body style="font-family:Tahoma">
                <p>Interfaces list:</p>
                    <xsl:for-each select="doc/members/member">
                        <xsl:sort order="ascending" />
                        <xsl:value-of select="fn:matches(., 'P\..+')" />
                        <br />

Why does I'm getting error:

Namespace does not contain any functions >

Where am I wrong? I found such code in examples, including!

Best Solution

In case you're performing the transformation with Visual Studio X, where X is not greater than 2008, this would be processed by an XSLT 1.0 processor (.NET's XslCompiledTransform or XslTransform). XSLT 1.0 uses XPath 1.0, not XPath 2.0 and its F & O (Functions and Operations), which only became a W3 Recommendation last year.

You have two options:

  1. Use a compliant XSLT 2.0 processor. If you prefer to stay within the .NET platform, then a suitable choice is Saxon.NET

  2. Just use the XPath 1.0 function starts-with(), which is sufficient to solve the current problem.
    The expression: starts-with(., 'P:Interfaces') is evaluated to true() if the string value of the context node starts with the string 'P:Interfaces' and to false() otherwise.

Another Xpath 1.0 function that may come handy for such type of processing is the function contains().

Xpath's 2.0 function ends-with() can be emulated in XPath 1.0 in the following way:

ends-with(s1, s2) ====substring(s1,string-length(s1)-string-length(s2)+1)=s2

where "===" means is "equivalent to".

Here we also used the XPath 1.0 functions substring() and string-length().