>Thanks for the reply. The ending condition can be any length. That's why I thought about using a UDF to rearrange to string. Even if it is not optimizable, shouldn't it take less time with the index rather than rushmore creating an index on the fly?
>
>

Rushmore most likely will not create any index for that type of comparison. It will have to process the entire table and run one comparison at a time. It does the same thing when an index exists and is not Rushmore optimzable. It just isn't used and the "brute force" technique is used.