No it shouldn't contradict what I said.

First you need to select the A nodes that have B elements with @a
containing 'foo'.  (Everyone provided methods that can do this.)

Second, you then copy or recreate the A nodes but only copy/reproduce
the B sub-nodes that have @a containing 'foo'.  (I think others will
agree with this, but neglected to see this as a requirement in your
first message.)

Darcy
On Wed, Sep 3, 2008 at 2:19 PM, mark bordelon <markcbordelon@xxxxxxxxx> wrote:
> I did understand you mean the first node, not the first attribute.
>
> Comprehension check: Does this at all contradict what Darcy wrote? I don't think it does...but just making sure.
>
> --- On Wed, 9/3/08, Andrew Welch <andrew.j.welch@xxxxxxxxx> wrote:
>
> From: Andrew Welch <andrew.j.welch@xxxxxxxxx>
> Subject: Re: [xsl] xslt 1.0 vs xslt 2.0 problem
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Date: Wednesday, September 3, 2008, 11:17 AM
>
> I wrote:
>
>> //A[contains((B/@a[1]), 'foo')]
>
> but meant:
>
> //A[contains((B/@a)[1], 'foo')]
>
> ...but hopefully you got the idea!
>
>
>
>
> --
> Andrew Welch
> http://andrewjwelch.com
> Kernow: http://kernowforsaxon.sf.net/