Thank you both for your time and suggestions. Yes, I stumbled in my test
across the white-space node and had to insert the strip-space
instruction for getting the desired results.
Ken's suggestion with normalize-space() is of course more elegant, being
independent of context and configuration.


Am Donnerstag, den 01.11.2007, 14:37 +0000 schrieb Andrew Welch:
> > Yes, I'm quite sure.  It was obvious by visually inspecting the
> > sample input XML ... there are so many text nodes there.  Did you run
> > your own suggestion as a test?
> 
> I did... but for some reason the -sall switch was set for Saxon, and
> my brain wasn't working, so I just trusted the output.
> 
> > I was only trying to help.  If I've misunderstood the original
> > poster's question, then I withdraw my suggestion and I apologize.
> 
> no I should apologize Ken, this is such a basic thing... for some
> reason at that point I thought whitespace text nodes wouldn't be
> matched by text() - it dawned on me soon after :-/
> 
> I've been learning the JPA these past few days, that's my excuse...