[xsl] How to create a node set that excludes some descendant elements?

[Rush Manbert <rush@xxxxxxxxxxx>]

Hi all,

My first post here and I want to start by saying how much I appreciate 
the big FAQ, the Jeni site, etc. It has all helped me tremendously.

I can't find an answer to this one, though, so here goes...

My XML doc has this basic structure:
<a>
 <b>
   <c>
     <!-- This is the section of interest -->
   </c>
 </b>
</a>

The <c> element can contain any combination of elements <d> through <z>. 
Elements <y> and <z> have special uses.

I want to create a global variable that contains the result tree 
fragment contained within element <c>, with the following restrictions:
I only want to include the first <y> element that is contained within 
<c>, no matter where it occurs. There may be no <y> elements present.
I want to exclude all <z> elements that are contained within <c>, no 
matter where they occur. Again, there may be none present.

Later on in my stylesheet, I use exslt:node-set() on the variable and 
process the node set.

For instance, given this source:
<a><b><c>
 <d>
   <z>
   <g />
 </d>
 <q>
   <r>
     <y />
     <z />
   </r>
   <y />
 <q>
 <y>
</c></b></a>

I want the selection to contain this:
<a><b><c>
 <d>
   <g />
 </d>
 <q>
   <r>
     <y />
   </r>
 <q>
</c></b></a>

(<z> elements are gone, only the first <y> element remains.)

I have tried many variations on the select portion of the variable 
definition. I can filter the immediate children of <c>, OR the second 
level children, etc., but I can't seem to come up with anything that 
handles <y> and <z> appearing at any depth in the descendant tree.

I'm prepared to be humiliated by some obvious solution... Can anyone 
please help?

Thanks,
Rush