Hi Joerg,

> And my comments/questions to Jeni's stylesheet:
>
>  >><xsl:template match="a">
>  >>  <xsl:variable name="a" select="." />
>  >>  <xsl:for-each select="$file2/b[@* = $a/@*]">
>
> I thought of a filter too, but I didn't find a sensible one. The
> above one isn't completely logical to me. What exactly does it mean?
> Maybe "all <b>s, whose all attributes are in <a> and don't have a
> different value"?

It means "all <b>s that have an attribute that has the same value as
at least one of <a>'s attributes". It's not a very strong filter, but
it means that if <a> is:

<a x="1" y="2" z="3" />

then you won't process <b> if it's:

<b x="4" y="5" />

but would process <b> if it was:

<b a="2" />

(and then discard it in the xsl:for-each).

Actually, given that the attributes on a have to be a subset of the
attributes on b (I think that's the way round it is now), then a
stronger filter would count the attributes as well:

   $file2/b[@* = $a/@* and
            count(@*) >= count($b/@*)]

There's a balance here between the amount of work that goes on in the
filter, and the amount of work that goes on within the xsl:for-each;
Zack should really test where the balance is to work out what filter
to use.
            
Cheers,

Jeni

---
Jeni Tennison
http://www.jenitennison.com/


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