Oliver,

>I'm not sure if the approach using key() is promising.

I had half suspected that key() would be more problematic for a generic
solution.

However, I'm currently using your updated stylesheet on a 800 node test
file.  At the moment it's performed about 10% in 12 mins.

Am I right in thinking that key() is more efficient than searching through
nodes using a <xslt:for-each>.  If so, then for large problems, key() may
be the only reasonable solution.

> However - it's an interesting idea. Nobody will stop you finding an 
> alternative approach! :-)

I'd like to claim credit for the idea, but it was Jeni Tennison's (another
one for her FAQ?).

Although I'm not looking for something super-fast, 2+ hours is a bit too
long. I'm resigned to thinking that a solution that uses a priori knowledge
about the format of the XML is my only way forward.

Thanks

Tom

-----------------------------------------------------------------
        Visit our Internet site at http://www.reuters.com

Any views expressed in this message are those of  the  individual
sender,  except  where  the sender specifically states them to be
the views of Reuters Ltd.


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