Yah, if you have room to load the whole doc, then Jaxen can select those
nodes for you with no problem.
Otherwise, I'd use dom4j's ElementHandler, and register to "/x/y/z" and
then maintain your own counter in your handler to do magic upon
every 2nd <z> maybe.
If you are writing SAX code, chances are you're not doing it the easy way.
-bob
On Fri, 26 Jul 2002, Rasmussen, John wrote:
> Yes, I did not phase that as well as I could have.
>
> My process receives an xml doc, and uses a xpath expression i.e. x/y/z[2] to
> designate a node on which to perform a pagination operation. I am
> proceeding with a jaxp sax class that will walk through the doc and build
> the hierarchal node name in order to test against the xpath expression.
>
> A bit tedious. If there were an easier, more efficient process I like to
> know about it.
>
> Thanks
>
> John
>
> -----Original Message-----
> From: bob mcwhirter [mailto:]
> Sent: Friday, July 26, 2002 9:58 AM
> To: Rasmussen, John
> Cc:
> Subject: Re: [xml-dev] sax parser and xpath
>
>
> > I need to write a java parser, sax I assume, that performs operations on a
> > node passed as an xpath expression.
> >
> > Anyone know if this is possible without hand-coding between startElement &
> > endElement ?
>
> I'm not certain I understand the question, but his might help:
>
> http://dom4j.org/
> -- Has an ElementHandler interface for processing
> sub-trees in a sax-like manner. Integrate jaxen.
>
> http://jaxen.org/
> -- An xpath engine. Integrated into dom4j, and compatible
> with jdom, dom, exml, and dom4j.
>
> -bob
>
--
Bob McWhirter
The Werken Company http://werken.com/
