On Mon, Nov 17, 2008 at 08:53:31PM -0000, Michael Kay wrote:

I quite like the fact that this generalizes.

$vSeq[index-of($vSeq,.)[$i]]

produces a list of the (distinct) values in $vseq that have at least $i
occurrences.

I'm tempted to prefer something with count(), for clarity,
but Dimitre's solution certainly does have the brevity
part down well :-)

I think this warrants a 'Dimitre' technique label?
   Up there with Steves IMHO.
For those with Mikes book, his notes explain the 'expansion'
that makes sense as iterating over the sequence, hence
allowing the use of the 'index'.

--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk