> I put an example below. So I don't know if this is
> going to be possible.  If anyone has any other thoughts, either for
> the first problem as an intellectual exercise, or to the 'revised'
> situation--including using XSLT or XQuery--that'd be great. Otherwise
> thanks for all the input!

Of course, this is possible. Below is one XPath expression the evaluation of 
which produces the wanted result:

    for $name in distinct-values(/*/*/*/author)
      return
         (concat('Books by ', $name, ': ',
           string-join(
                   /Library/Branch/book
                   /(if(author=$name)
                       then concat(position(),
                                   if(following::author[. eq $name])
                                      then ','
                                      else ''
                                   )
                             else ()
                     ),
                     ''
                  ),
           '
'
          )
         )



When evaluated with the current document the one supplied in your last 
message, the result is:


Books by John: 1,3,5
 Books by Mary: 2,6
 Books by Mark: 4



Cheers,
Dimitre Novatchev


<nikoromano@g...> wrote in message 
news:19b702bd-1661-488d-b7fb-715a8b27172e@d......
> Hi Guys
>
> Thanks for all the input. Dimitre, your suggestion is something I had
> tried as well. Unfortunately I don't think it quite works... Well,
> more specially it *does* work for the one example I posted, but
> imagine if I added another book below the last one with the author
> Mary. (I put an updated XML snippet below) The the output for John is:
> 1, 3, 5,
> (Since the 5th book is no longer the last book overall, a comma is
> added after the 5.)
>
> The reason I am trying to use XPath is I am being forced to use Altova
> StyleVision for this project, and as far as I am aware,  StyleVision
> won't let me do this any other way (i.e. using the XLST or XQuery).
>
> After I posted this, I realized the problem is actually more
> complicated than I had previously thought. The point is actually to
> create a detailed list of books, sorted by title (title is just
> another element inside book) with a summary on top telling which books
> below correspond to the given authors by the position the book appears
> in the report. I put an example below. So I don't know if this is
> going to be possible.  If anyone has any other thoughts, either for
> the first problem as an intellectual exercise, or to the 'revised'
> situation--including using XSLT or XQuery--that'd be great. Otherwise
> thanks for all the input!
>
> Output example
> Books by John: 1, 2, 5
> Books by Mary: 3, 6
> Books by Mark: 4
>
> Book 1:
> Title: A Book Title Starting With 'A'
> Author: John
>
> Book 2:
> Title: Books Are Fun!
> Author: John
>
> Book 3:
> Title: Can You Tell Me How To Get To Sesame Street?
> Author Mary
> ...
>
>
>
> XML Example
> <?xml version="1.0" encoding="UTF-8"?>
> <Library>
> <Branch>
> <book>
> <author>John</author>
> </book>
> <book>
> <author>Mary</author>
> </book>
> </Branch>
> <Branch>
> <book>
> <author>John</author>
> </book>
> <book>
> <author>Mark</author>
> </book>
> <book>
> <author>John</author>
> </book>
> <book>
> <author>Mary</author>
> </book>
> </Branch>
> </Library>
>
>
> On Dec 2, 10:26 pm, "Dimitre Novatchev" <dimit...@tpg.com.au> wrote:
>> "Joe Kesselman" <keshlam-nos...@comcast.net> wrote in message
>>
>> news:gqmdnXmD4cUuFszanZ2dnUVZ_tLinZ2d@c......
>>
>> >>           /(if(author='John')
>>
>> > I believe if/then/else expressions require XPath 2.0 or XQuery. If 
>> > that's
>> > what you have available, go for it;
>>
>> Yes, this was the code in the OP first post.
>>
>> Cheers,
>> Dimitre Novatchev
>