No kidding! I have a 220 record subset of the data I use for testing. About the only place that I am inefficient, I choose to take as a performance hit for better designed code. I have each different table that I am writting to split into its own subroutine. That way, I only have to debug one procedure if there is a problem, and I don't have to worry about leaving variables defined when I am done with it.

>>Thanks you guys!
>>
>
>Oh, in that case George is right, no difference...I'd look for improvements elsewhere. But you know, I have apps that take much longer than that, so don't feel too bad...
>
>You might run it through the the coverage log, but with a very small data set/number of loops, etc., to see if anything shows up that might help you.
>
>Don't run it with coverage log as-is, though, or you'll be waiting until next year... < s >