I am trying to work out a bug we've found during our last iteration of testing. It involves a query which uses a common table expression. The main theme of the query is that it simulates a 'first' aggregate operation (get the first row for this grouping).I am trying to work out a bug we've found durin