- From: David Carlisle <davidc@nag.co.uk>
- Date: Tue, 3 Aug 2004 23:40:46 +0100
- To: chamberl@almaden.ibm.com
- Cc: public-qt-comments@w3.org
It's not clear whether the xpath1 distinction between predicates in axis
steps and filters is really doing much in xpath2/xquery but anyway as it
is, you are right (that I'm wrong:-)
So It seems easiest way to get a filter expression is to use a ()
expression rather than an axis step; something like
fn:doc("zoo.xml")/something/(dog|cat)[1]
> It's not easy to find a reasonable example of a filter expression inside a
> path expression!
hmm I'm not sure my example meets a reasonableness criterion:-)
why can't ForwardStep | ReverseStep be moved into PrimaryExpression so
that AxisStep can be deleted and filterexpression unified with
stepexpression
that would unify all uses of [...] into one place (but this is gut
reaction while answering your mail, I haven't really analysed the
grammar that would result.
David
________________________________________________________________________
This e-mail has been scanned for all viruses by Star Internet. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________
Received on Tuesday, 3 August 2004 18:41:06 UTC