Yes, there are more levels of LovelyTags , and they can be follow in
random order and have no fixed length, i.e. all this paths should be
valid

/MyRoot/LovelyTag1/LovelyTag2/Item
/MyRoot/LovelyTag1/LovelyTag2/LovelyTag2/LovelyTag1/Item
/MyRoot/LovelyTag2/LovelyTag1/LovelyTag2/LovelyTag1/LovelyTag2/Item

Ok I think my first idea with replacements (and latest "evolution" to
use external namespace object from C#, where more easily to detect such
issues) should be used.

Martin Honnen wrote:
> centur@g... wrote:
>
> > It's a bit hard to explain question in plain English or plain code, I
> > try to provide simple
> > XML
> > <MyRoot>
> > <LovelyTag1>
> >   <LovelyTag2>
> >    <Item/>
> >   </LovelyTag2>
> > </LovelyTag1>
> >
> > <LovelyTag1>
> > <BadTag1>
> >   <LovelyTag2>
> >    <Item/>
> >   </LovelyTag2>
> > </BadTag1>
> > </LovelyTag1>
> > </MyRoot>
> >
> > Is there any way to evaluate query:
> > "I want to get all items with only LovelyTag1 and LovelyTag2 as
> > ancestors (and myroot of course) but items with any parents except
> > LovelyTag1 and 2 should be discarded".
>
> Well for your example
>    /MyRoot/LovelyTag1/LovelyTag2/Item
> should do.
> If there are more and arbitrary levels then it is not that easy.
>
> 
> -- 
> 
> 	Martin Honnen --- MVP XML
> 	http://JavaScript.FAQTs.com/