>Hi Fabio,
>
>Nice, that we saw the same useless code...
>
>Well, if the index is not temporally, but built over time with the work on that table, that time won't hurt. But even with the time needed to build the index I came out with an average of 1.28 secs needed by SQL and 0.92 secs by the scan. This divides into 0.89 secs for building the index (SET TALK OFF!) and 0.13 secs for the scan. And the times needed had small variations with several (10) test runs.
>
>Bye, Olaf.

Olaf,
the current SQL engine is not 100% efficient.
I'm sure that a 2x improvement is possible into the current VFP engine.