"Passin, Tom" <tpassin@xxxxxxxxxxxx> wrote in message
news:5D3C2276FD64424297729EB733ED1F7601D1BD10@xxxxxxxxxxxxxxxxxxxxxx
> [Dimitre Novatchev]
> >
> > It will also "work" for more than one node (in case $tagPath
> > evaluates to a
> > nodeset of two or more nodes) and most probably is not what
> > the author of
> > the original message wanted.
> >
>
> I think I may have misunderstood here, and perhaps you misunderstood me
> as well...  I understood the original question to ask for matching a
> __string__ containing a path expression, like "/a/b/c" (that is the kind
> of example O thought that Yue Ma showed).  For my own post, I put the
> path string into a variable, thinking that it would probably eventually
> be sent in via a parameter.
>
> Re-reading your comment, I notice that you might have  thought that I
> was selecting a node set into the variable, instead of a string, but
> that is not the case.

No, I also understand that the variable contains a string.

The problem of the expression (1):

/a/b/c

being generated for the node uniquly matched by (2):

/a/b[2]/c[3]

is that the former matches more than one node.

Whoever wants to identify a single node will not use (1) , but an expression
equivalent to (2).

This is why the chances of

"/a/b/c"

to be equal to the (string) expression identifying uniquely a node -- these
chances are close to zero.


I think that the "standard" answer is to use a xx:evaluate() extension
function.

It is also good to ask why such dynamic evaluation is necessary in the first
place.



=====
Cheers,

Dimitre Novatchev.
http://fxsl.sourceforge.net/ -- the home of FXSL




 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list