Looking for good example...

doxley · June 15th, 2009, 12:08 PM

I have a problem that I'm certain has been solved and probably has an elegant solution, but after much looking, I can't find it. I'm hoping someone can give me a pointer.

I need to process an xml document with XSLT such that I use one mode for all nodes that precede some identifiable node within the document and an a different mode for all subsequent nodes. IOW, I want to partition the tree into two parts and process each with a different collection of templates.

I have a relatively straightforward solution, but it relies on my providing an explicit value for a position in order to construct the predicate to separate the two halves.

It seems so simple, I'm sure someone has an elegant solution.

Thanks for your help.

--Don

Martin Honnen · June 15th, 2009, 12:23 PM

Can you show us a sample input and your current solution? Then it might become clearer what you are trying to achieve.
Also tell us whether you want to use XSLT 2.0 or 1.0.

doxley · June 15th, 2009, 04:31 PM

Martin,

I'm actually working with XSLT 2.0, so a solution depending on new features is fine.

My current code l is shown below. It does not operate exactly as I would like, though the difference is small. It doesn't use a mode for the first part of the document and I have much more knowledge of the tree elements than I would like in the base template.

The template is actually parsing a modified Word 2007 xml document though I've done a good deal of elimination of elements and changed the namespace prior to invoking this template. This is a stage in which I'm applying some corrections for poor style choices on the part of the document author in preparation for parsing the document itself.

Hope this helps - thanks for suggestions.

--Don

Code:

<xsl:stylesheet version="2.0"
          xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
          xmlns:pkg="http://schemas.microsoft.com/office/2006/xmlPackage"
        xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main"
        xmlns:g="gptg:PMSA:gptg.net">

  <xsl:output method="xml" omit-xml-declaration="no" indent = "yes" encoding = "utf-8"/>
  <xsl:strip-space elements="*"/>
  <xsl:namespace-alias stylesheet-prefix="w" result-prefix="g"/>
  <xsl:namespace-alias stylesheet-prefix="pkg" result-prefix="g"/>


  <!-- Default template - handles first part of document and serves as a default mode -->
  <!-- **the hard coded position is one of the issues - it corresponds to the node that is splitting the document -->
  <xsl:template match="node() | @*">
    <xsl:copy>
      <xsl:apply-templates select="@*"/>
      <xsl:apply-templates select="g:document | g:body | g:paraStyle | g:para[position() &lt; 4] |g:text | text()"/>
    </xsl:copy>
  </xsl:template>

  <!-- template to determine a proper mode for the back portion of the document.   -->
  <!-- **the hard coded values correspond to the location of this node which should really be identified as the node which
       satisfies the match condition -->
  <xsl:template match="/g:document/g:body/g:para[substring(normalize-space(g:text[1]),1,7)='Section']">
    <xsl:variable name="protocol" select="normalize-space(substring-after(substring-before(g:text[1],':'),'Section'))"/>
    <xsl:choose>
      <xsl:when test="$protocol='4'">
        <xsl:apply-templates select="/g:document/g:body/*[position() &gt; 2]" mode="Pro4"/>
      </xsl:when>
      <xsl:otherwise>
        <xsl:apply-templates select="/g:document/g:body/*[position() &gt; 2]" mode="None"/>
      </xsl:otherwise>
    </xsl:choose>
  </xsl:template>

  <!-- copy template in case there is no real work to be done to the back portion of the document -->
  <!--  ** this template is called out separately for purposes of later extension  -->
  <xsl:template match=" node() | @*" mode="None">
    <xsl:copy>
      <xsl:apply-templates select="@*" mode="None"/>
      <xsl:apply-templates mode="None"/>
    </xsl:copy>
  </xsl:template>

  <!-- copy template for the modal templates that work on the back portion of the document -->
  <!--   ** there are many templates matching specific conditions in this mode -->
  <xsl:template match="node() | @*" mode="Pro4">
    <xsl:copy>
      <xsl:apply-templates select="@*" mode="Pro4"/>
      <xsl:apply-templates  mode="Pro4"/>
    </xsl:copy>
  </xsl:template>

</xsl:stylesheet>

Martin Honnen · June 16th, 2009, 09:40 AM

It has not become much clearer to me what you want to achieve.

XPath 2.0 has the operators << and >> which you can use to check whether node1 is before node2 in document order (<<) or the other way round (>>) so maybe you want e.g.

Code:

<xsl:variable name="split-el" select="/root/foo/bar"/>
<xsl:apply-templates select="/g:document/g:body/*[. &gt;&gt; $split-el]" mode="Pro4"/>

to process those nodes coming after $split-el in mode 'Pro4'.