Thanks for the info - it's a little frustrating to know that doing things the "right" way, don't always lead to the best solutions. I guess that's why the UT has done so well. I really appreciate the experience that those of us who are considered novices can tap into.

>Howard,
>
>If you want to verify the optimization of a query like this (using sys(3054))
>
>select * from sometable where upper(somefield) = "somevalue" into cursor somecursor
>
>You'd have to have an index tag on upper(somefield) (but you knew that right?)
>
>having SET DELETED ON will report partial optimization. This may lead you to suspect your expression and indexes don't match. If you SET DELETED OFF (just to test optimization) and run the query you can be sure you're getting "Full Rushmore Optimization" for each situation.
>
>>>Having an index on DELETED() will help you achieve "Full Rushmore Optimization", by definition. But "Full Rushmore Optimization" is not always optimal. Please read my recent entry in FAQ for a detailed explanation.
>>>