On the subject of the expensive sort in the execution plan I had a note back from Bob Beauchamp saying that there is an internal stream aggregator operation that causes the sort and unfortunately there is no way around it other than to put the clustered index on the column(s) used in the GROUP BY statement.:crying: