I suppose I could do that. I have not tried that approach before. I suspect in this case that all the trimming and padding I'm doing defeats the indexes, making this a long running prospect, but I fear getting less accurate results.

My brain is a bit fuzzy today, but I wonder if there's a definitive way to test that the results are identical between 2 queries, if I want to test various approaches? If I get a result cursor with 1100 rows, can I quickly prove that results remain identical?

>Can you just build your select as a string (using textmerge) and then execute using macro substitution? This way you would not need such complex logic.
>