Sorry. No new ideas on what the problem might be. I'm sure you have a good reason for working with an array rather than a cursor so I won't suggest re-working the application to get rid of the array. How about something like copying the record subset to a temp dbf and creating the array from there? Don't know why that should be any better, but you never know. Good luck.

>>COPY TO ARRAY has problems with Memo fields. I don't have a copy of FPD 2.5 at hand but in FPW2.6 what I "THINK" happens is that all memo fields are moved to the last columns in the array and are populated with .F. (this is what happened with my first experiment on this).
>>
>>If you send me the table I may be able to find a pattern for you when I get back home to an FPD archive.
>>
>>HTH............Rich
>
>There aren't any memos involved at all, the array only ever has character fields and one date. It's not a constant happening, maybe 1% of the time (or my users would be killing me), and happens to at least two tables and copy statements. One of the tables is 16 MB (333K records) with ~50+% deleted, and the COPY TO statment has conditions that filter it down to the expected half-dozen or so. They're properly Rushmore optimized and don't appear to be corrupted. Any other ideas?
>
>Chris.