Well Dave, I was hoping for some real benchmark data, but thanks anyway.

Of course the sql needs to be changed to the dialect specific to the platform - that is always a given. I'm not sure why you have to use CASE on sql server, but then I am a sql server newbie (not an sql newbie just a sql server newbie). Perhaps that is what makes it soooo slow and the syntax so ugly. On Access you have to use the HAVING clause for this to work but the boolean calculation is only a very small performance hit. On MySQL v 4 other problems arise, like the lack of table-type subquery support. v5 on the other hand would allow the sub-query select so that is just fine.

As I said, if all else fails then my method will work on any platform, just adjust the sql to the local dialect. If the platform has other tools then use them.