Thanks for the hint, I'll work on the perfomance then and give your
previous recommendation a closer look.
Andrew Welch wrote:
<xsl:template match="*[(_at_)index]">
<xsl:copy>
<xsl:copy-of select="@*[not(name()='index')]"/>
<xsl:variable name="p" select="count(preceding::*[(_at_)index])+1"/>
This is fine if it does what you need, but you should be aware that
using the preceding::* axis in this way leads to O(n^2) performance.
It's much better practise to store the data you need in a nodeset, and
then query that. This means you will do one pass for the nodeset, and
one pass for the data regardless of the number of elements in your
source.
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
--
Mit freundlichen Grüßen / With kind regards
Volker Witzel
____________________________________________
IBM Business Consulting Services
Karl-Arnold-Platz 1a, D-40474 Düsseldorf
fon (0211) 476-1924 (intern *129-1924)
mobil 0172-73 66 586
mail vwitzel(_at_)de(_dot_)ibm(_dot_)com
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list