>>Yes. The program first deletes some records from ndxfl1 (none in this case), then uses SQLEXEC to populate another cursor. There is some code in between I better trace through before taking up any more of your time. Will get back to you. Thanks for your help and dedication, as always.
>
>Any time.

Well, here is the end of this day long saga, at least insofar as it concerns the general community. There is not a bug in VFP, at least I don't think so. We have a goofy little routine that massages the SQL command string and is somehow messing things up. I don't know why because all it uses is basic, ancient string functions like SUBSTR() and ATC(). The guy who wrote this stuff avoided comments like the plague so it isn't even clear to me yet what the routine is trying to do. And I don't know why this suddenly cropped up under VFP 9, using the exact same code and data. Somehow the combination of one line ending with no spaces, just a linefeed, and the next line starting with no spaces, sets it off. It winds up pulling the first word in the second line out of the string.

If this turns out to be due to a change in VFP behavior I will follow up with another message. Thanks again for your help.