For those who wonder about VFP performance on a back end database (SQL Server) Here is a recent problem I have had. When committing data to the backend, I have all of my sql passthrough statements in an array (plus the tableupdates) so that I can rollback or commit depending on whether everything is okay.

I had problems because some of the sql passthrough statements that call a stored procedure to update a table were not working correctly. The problem was traced to the fact that multiple stored procedure calls, using SQL passthrough, were creating records with the SAME datetime stamp. Right down to the 1000th of a second. I had to modify the stored procedure that made the assumption that the datetimes were different.

I think it would be hard to get better performance than that.