Hi Manousos,
  Please try this XSL. It uses nodeset extension
function.

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:xalan="http://xml.apache.org/xalan"
exclude-result-prefixes="xalan">

<xsl:output method="xml" indent="yes"/>
	
<xsl:variable name="root" select="/" />
	
<xsl:template match="/ROWSET">
  <set>
     <xsl:variable name="rtf">
       <xsl:for-each
select="descendant::ELEM_INFO_ITEM">
         <xsl:if test="((position() mod 3) = 1) and
(position() &gt; 1)">
           <x>
             <xsl:value-of select="." />
           </x>
         </xsl:if>
       </xsl:for-each>  
     </xsl:variable>
	    
     <xsl:for-each select="xalan:nodeset($rtf)/x">
       <xsl:variable name="a" select="." />
       <xsl:if test="position() &lt; last()">         

         <xsl:variable name="b"
select="following-sibling::x[1]" />
           <interior>	          
	     <xsl:copy-of
select="$root//ORDINATES_ITEM[(position() &gt; ($a -
1)) and (position() &lt; $b) ]" />
	   </interior>
	</xsl:if>  
	<xsl:if test="position() = last()">
	   <interior>
	     <xsl:copy-of
select="$root//ORDINATES_ITEM[position() &gt; ($a -
1)]" />
	   </interior>
	</xsl:if>
      </xsl:for-each>	    
  </set>  
</xsl:template>
	
</xsl:stylesheet>

Regards,
Mukul

--- mankar@xxxxxxxxxxxxxxx wrote:

> Hallo,
> 
> Let me please rephrase my question concerning the
> recursive splitting  of a
> particular node set, in order to make it less
> obscure.
> 
> I have this source xml
> 
> <?xml version="1.0" encoding="ISO-8859-7"?>
> <ROWSET>
>       <ROW num="1">
>             <SHAPE>
>                   <ELEM_INFO>
>                        
> <ELEM_INFO_ITEM>1</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>1003</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>50</ELEM_INFO_ITEM>
> 
>                        
> <ELEM_INFO_ITEM>3</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>2003</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>50</ELEM_INFO_ITEM>
> 
>                        
> <ELEM_INFO_ITEM>8</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>2003</ELEM_INFO_ITEM>
>                        
> <ELEM_INFO_ITEM>50</ELEM_INFO_ITEM>
>                   </ELEM_INFO>
>                   <ORDINATES>
>                        
> <ORDINATES_ITEM>52</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>45</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>54</ORDINATES_ITEM> position 3
>                        
> <ORDINATES_ITEM>57</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>76</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>78</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>56</ORDINATES_ITEM>
>                        
> <ORDINATES_ITEM>23</ORDINATES_ITEM> position 8
>                        
> <ORDINATES_ITEM>32</ORDINATES_ITEM>
>                   </ORDINATES>
>             </SHAPE>
>       </ROW>
> </ROWSET>
> 
> I want to use the ELEM_INFO_ITEM[position() mod 3 =
> 1][position() &gt; 1]
> element's values, which are for this particular
> example 3 and 8, as the
> appropriate positions to split the ORDINATES_ITEM
> node set into two
> separate  nodes-sets.I mean select ORDINATE_ITEM
> elements from position 3
> to 7 as children of an interior element and
> ORDINATE_ITEM elements from
> position 8 to the last as children of another
> interior element.
> The desired output should look like:
> 
> <set>
> <interior>
> <ORDINATES_ITEM>54</ORDINATES_ITEM>
> <ORDINATES_ITEM>57</ORDINATES_ITEM>
> <ORDINATES_ITEM>76</ORDINATES_ITEM>
> <ORDINATES_ITEM>78</ORDINATES_ITEM>
> <ORDINATES_ITEM>56</ORDINATES_ITEM>
> </interior>
> <interior>
> <ORDINATES_ITEM>23</ORDINATES_ITEM>
> <ORDINATES_ITEM>32</ORDINATES_ITEM>
> </interior>
> </set>
> 
> 
> This is easy to manage, when the source xml is
> always the same, but it is
> not.It is generated dynamically and the total number
> of ELEM_INFO_ITEM
> elements is different each time.This number is
> always multiple of three, (9
> in this example) and cannot be less than six.This
> means that the number of
> the generated interior elements also varies and
> depends on the number of
> ELEM_INFO_ITEM elements.
> 
> To be more specific i want, depending on my source
> xml, to generate as many
> interior elements as the
> ELEM_INFO_ITEM[position()mod3=1][position()&gt; 1]
> are, and i get to do this using something
> like:
> 
> <xsl:for-each
>
select="ROWSET/ROW/SHAPE/ELEM_INFO/ELEM_INFO_ITEM[position()mod3=1][position()
>  &gt;1]">
> <interior>...</interior>
> </xsl:for-each>
> 
> If i for example, had another source xml identical
> with the first but with
> the three last ELEM_INFO_ITEM elements missing, the
> previous code would
> generate exactly one interior element, as i wish it
> to be.
> 
> The hard part is to select the appropriate
> ORDINATE_ITEM nodes and pass
> them as children to the relevant interior
> element.Can i automate this task
> through xslt code?
> I mean, since the source is not always the same how
> can i recursively
> specify that the first interior element should
> always contain ORDINATE_ITEM
> elements whose position is greater than or equal to
> ELEM_INFO_ITEM[position()mod3=1][2] and less than
> ELEM_INFO_ITEM[position()mod3=1][3],
> and the second interior element should always
> contain ORDINATE_ITEM
> elements whose position is greater than or equal to
> ELEM_INFO_ITEM[position()mod3=1][3] and less than
> ELEM_INFO_ITEM[position()mod3=1][4] etc.
> 
> How can i do this, since i can't know beforehand,
> for example if an element
> ELEM_INFO_ITEM[position()mod3=1][4]exists, as this
> always depends on my
> source xml.Is there a way to recursively define the
> appropriate ceiling and
> threshold ORDINATE_ITEM position for each sequential
> interior element.
> 
> Please,let me know if  there is a way to obtain
> this.
> 
> 
> Regards
> Manousos
> Athens



		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail