>>>
>>>Wouldn't that benefit the GROUP BY?
>>>
>>>
>>>No, GROUP BY, HAVING, ORDER BY are not optimizable.
>>
>>Why is that?
>
>AFAIK, they are applied to the resulting cursor after all the WHERE and JOIN clauses have been processed. There are no indexes available for optimization to take place.

I have no idea of how the internal workings of Rushmore are constructed, but I was under the impression that if an index existed, a temporary one may be built on the result set. Again, I have no idea if that is how it works or not.