Could you maybe tell me if that new query does the job? I really think this is a good general purpose workaround for that limitation. Maybe even other databases that don't support that could benefit from that. DB2 for example also doesn't support tuple distinct counts and I guess there are even more.