Hi Kevin,
I now know I mean @@RowCount. In VFP, with native tables, my current GetTheNextKey function doesn't use record locking either. I'm using a cursorAdapter with ConflictCheckType = 3. If there is a conflict writing the new key to the keys table, I know that someone else has updated it since I got the last key, so I must retry. When I run two processes in a loop, both asking for the next key, I actually get better performance than when I use conventional locking.

>Thanks, Venelina (and everyone else)
>
>I got side-tracked on a bunch of different issues, and am now coming back to this one. Venelina and Charlie's responses in particular were interesting, and I'm evaluating every idea right now. Sorry for not responding sooner.
>
>Thanks!
>Kevin