RE: [xsl] line numbering

Subject: RE: [xsl] line numbering
From: "Andrew Welch" <ajwelch@xxxxxxxxxxxxxxx>
Date: Thu, 5 Aug 2004 15:17:05 +0100
> Hi there,
>
> I have some poems marked up something like:
>
> <body>
> <div type="poem">
> <lg type="stanza">
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> </lg>
> <lg type="stanza">
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> </lg>
> <p>This is something not counted as a line</p>
> <lg type="stanza">
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> <l>This is a line of verse</l>
> </lg>
> </div>
> <!-- and multiple poem div's like this here... -->
> </body>
>
> What I want to end up with is for each line to be
> given an @id in the html output like:
> <span id="poem3line10">This is a line of verse <span
> class="number">10</span></span>
>
> <xsl:template match="l">
> <xsl:variable name="linenumber"
> select="count(preceding-sibling::l)+1"/>
> <span class="line"
> id="line{$linenumber}"><xsl:apply-templates /></span> </xsl:template>
>
> But now I need to account for two factors:
> a) that the linenumbering needs to take account of
> the existence of the line-groups (lg) and that it
> should only count back within it's own <div> ancestor.
> b) that I want to number the lines every 5 lines.
>
> Is position() a better way to do this than counting the
> preceding-sibilings?

If you have a large input xml you may find counting preceding siblings
to be expensive, and that its better to do the numbering in a variable
in an initial pass of the xml.  For example using this stylesheet:

<xsl:stylesheet version="1.0"
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
  xmlns:exsl="http://exslt.org/common";
  extension-element-prefixes="exsl">

<xsl:output indent="yes"/>

<xsl:key name="lines" match="line" use="@id"/>

<xsl:variable name="lines-rtf">
  <xsl:for-each select="/body/div">
    <poem number="{position()}">
      <xsl:for-each select="lg/l">
        <line id="{generate-id()}" number="{position()}"/>
      </xsl:for-each>
    </poem>
  </xsl:for-each>
</xsl:variable>
<xsl:variable name="lines" select="exsl:node-set($lines-rtf)"/>

<xsl:template match="/">
 <div>
   <xsl:apply-templates select="/body/div/lg/l"/>
 </div>
</xsl:template>

<xsl:template match="l">
  <xsl:variable name="genId" select="generate-id()"/>
  <span>
    <xsl:attribute name="id">
      <xsl:for-each select="$lines">
         <xsl:for-each select="key('lines',$genId)">
           <xsl:value-of
select="concat('poem',parent::poem/@number,'line',@number)"/>
         </xsl:for-each>
      </xsl:for-each>
    </xsl:attribute>

    <xsl:value-of select="."/>

    <xsl:if test="position() mod 5 = 1">
      <span class="number">
        <xsl:for-each select="$lines">
          <xsl:value-of select="key('lines',$genId)/@number"/>
        </xsl:for-each>
      </span>
    </xsl:if>
  </span>
</xsl:template>

</xsl:stylesheet>

With your example xml gives this result:

<div>
   <span id="poem1line1">This is a line of verse<span
class="number">1</span>
   </span>
   <span id="poem1line2">This is a line of verse</span>
   <span id="poem1line3">This is a line of verse</span>
   <span id="poem1line4">This is a line of verse</span>
   <span id="poem1line5">This is a line of verse</span>
   <span id="poem1line6">This is a line of verse<span
class="number">6</span>
   </span>
   <span id="poem1line7">This is a line of verse</span>
   <span id="poem1line8">This is a line of verse</span>
   <span id="poem1line9">This is a line of verse</span>
   <span id="poem1line10">This is a line of verse</span>
   <span id="poem1line11">This is a line of verse<span
class="number">11</span>
   </span>
   <span id="poem1line12">This is a line of verse</span>
</div>

(I'm not sure if the line numbering 'every five lines' is correct, I
would guess you would want 1,5,10,15 and so on, so a bit more logic is
required there)

In the stylesheet all the numbering is done using the position()
function as the nested for-each's iterate over xml.  The variable is
then queried using a key to maximise performance when accessing the
variable.  This minimal stylesheet should be O(2n) complexity, counting
preceding-siblings would be slightly worse in this case.


cheers
andrew

Current Thread