Hello Michael,



it's really easy: use preceding-sibling and following-sibling axes.



<xsl:value-of select="preceding-sibling::node/@id"/>
<xsl:value-of select="following-sibling::node/@id"/>

Regards,



Joerg



Michael Peet wrote:
Hi Everyone,



With an input XML of:



<xml>
 <a>
   <b>
     <node id="1"/>
     <node id="2"/>
     <node id="3"/>
     <c>
       <node id="4"/>
       <node id="5"/>
       <node id="6"/>
       <d>
         <node id="7"/>
         <node id="8"/>
         <node id="9"/>
       </d>
       <e>
          <node id="10"/>
          <node id="11"/>
          <node id="12" current="true"/>
       </e>
     </c>
   </b>
   <b>
     <node id="13"/>
   </b>
   <b>
     <node id="14"/>
   </b>
   <b>
     <node id="15"/>
   </b>
 </a>
</xml>

I'm trying to identify those node elements immediately preceding and 
following the one marked current="true", but ONLY if within the same <b> 
element.  For example, if node 6 is marked current, 5 and 7 will be 
preceding and following, respectively.  If node 12 is current (as 
shown), 11 will be preceding, and since there are no more node elements 
within that <b>, there will be no following.



Note that the depth of any given node element may be considered 
arbitrary within the heirarchy; the exact structure will not be known.  
I also need to output something if there is no "preceding" or 
"following" node.  There will be only one node with the current="true" 
attribute.



I have come up with 2 approaches so far (following only shown as the 
same approach can be used with preceding as well):



<xsl:stylesheet version="1.0" 
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">



 <xsl:output encoding="ascii" omit-xml-declaration="yes" indent="yes"/>
 <xsl:strip-space elements="*"/>

 <xsl:variable name="current-node" select="//node[@current]"/>



 <xsl:template match="/">



   current: <xsl:value-of select="$current-node/@id"/><br/>



   <xsl:choose>

     <xsl:when 
test="$current-node/following::node[1]/ancestor-or-self::b//node[@current]"> 



       following: <xsl:value-of 
select="$current-node/following::node[1]/@id"/><br/>

     </xsl:when>

     <xsl:otherwise>

       no more<br/>

     </xsl:otherwise>

   </xsl:choose>



   <xsl:choose>

     <xsl:when test="generate-id($current-node/ancestor-or-self::b) = 
generate-id($current-node/following::node[1]/ancestor-or-self::b)">

       following: <xsl:value-of 
select="$current-node/following::node[1]/@id"/><br/>

     </xsl:when>

     <xsl:otherwise>

       no more<br/>

     </xsl:otherwise>

   </xsl:choose>



 </xsl:template>



</xsl:stylesheet>



The first approach tests whether the "following" node's <b> ancestor 
contains a node with the current="true" attribute.



The second approach tests if the <b> elements which are ancestors of the 
current and following nodes are identical via the generate-id() function.



Can anyone comment on the relative merits / problems of either 
approach?  The first approach uses the frowned-upon //.  Can anyone see 
a way to use a clever predicate instead of the ancestor-or-self axis?  
Is there an obvious solution that I'm missing?



Many thanks for your time,



Mike




XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list