Subject: Re: Catch ALL | Failed template rule
From: Stef <stef@xxxxxxxxx>
Date: Tue, 22 Mar 2005 15:20:00 -0500
|
On Tue, Mar 22, 2005 at 07:59:15PM -0000, Michael Kay wrote:
> > I guess I was looking for the
> > cleanest way to exercise this. I guess it would be this:
> > <xsl:apply-template select="not(foo/bar[@a='b'])"
> > mode="unmatchedresults"/>
>
> No, the result of the selection has to be a set of nodes, it can't be a
> boolean.
>
> The idea in your head of invoking a template rule if something isn't there
> is not a ridiculous one, but it's something that the language doesn't
> support, and couldn't be made to support without introducing some new
> fundamental concepts. So the best one can say is: it ain't going to happen;
> find another route to your destination.
>
Jumping in at the eleventh hour and probably am hugely
wrong here, but, wouldn't the following work ?
<xsl:apply-template select="*[not(foo/bar[@a='b'])]" mode="unmatchedresults"/>
This would match any node that did not have the foo/bar
entity with an attribute of a that was b. you would (of
course) need to provide some other search criteria such
as name()="corefoo" so that it didn't match elements
you were not interested in.
As a sidenote, I have used the following in my xslt to
match any sub-element that was -not- a prev, next or
alternativenames
<xsl:for-each select="*[not(name()='prev') and not(name()='next') and no
t(name()='alternativenames')]">
Is there any 'cleaner' or 'saner' way to write this ?
again, apologies if this is not what was asked for,
I am simply trying to offer something that has worked
for me in a somewhat similiar fashion.
regards
Steph
| Current Thread |
|
Stef - 22 Mar 2005 20:25:41 -0000 <=
Nicola Harlow - 24 Mar 2005 13:40:04 -0000
|
|