Stefan Franke wrote:


> The XML I'm using looks like this:
> 
> <SAMMLUNG
>    xmlns:buch="http://www.example.com/buecher"
>    xmlns:cd="http://www.example.com/cds">
> 
>    <buch:ARTIKEL Status="vorhanden">
>       <buch:TITEL>The Marble Faun</buch:TITEL>
>       <buch:AUTOR>Nathaniel Hawthorne</buch:AUTOR>
>       <buch:PREIS>10.95</buch:PREIS>
>    </buch:ARTIKEL>
>    <cd:ARTIKEL>
>       <cd:TITEL>Violinkonzerte 1, 2 und 3</cd:TITEL>
>       <cd:KOMPONIST>Mozart</cd:KOMPONIST>
>       <cd:PREIS>16.49</cd:PREIS>
>    </cd:ARTIKEL>
> </SAMMLUNG>
> 
> 
> If I want to display e.g. all the CDs of Mozart, I would do something like 
> this if I wouldn't have to deal with namespaces:
> 
> string(//*/TITEL[following-sibling::KOMPONIST='Mozart'])

Not really, if you apply the string function to a nodeset then you get 
the string value of the first node in document order so the above 
doesn't give you all CD titles but only one.

> My suggestion when dealing with namespaces would be something like this (but 
> it doesn't work, as there is no TITEL-element anywhere, but I don't know how 
> to get it in there):
> 
> string(//*[local-name()='ARTIKEL' and 
> namespace-uri()='http://www.example.com/cds']/*[string(local-name()='KOMPONIST') 
> = 'Mozart'])

How are you evaluating your XPaths? You should simply be able to declare 
prefixes for the namespaces and then use e.g.
   //prefix:TITLE
instead of going to the pain of using namespace-uri comparisons.
So I think you should better tell us about the software you are using to 
evaluate XPaths and then probably someone can tell you how to use 
namespace prefixes bound to namespaces when evaluating XPath expressions.

-- 

	Martin Honnen
	http://JavaScript.FAQTs.com/