I had a similar problem to your first scenario. Turns out there was validation code buried deep in the class hierarchy I inherited that was doing a SEEK() when it wasn't expected and landing on a random record.

It was downright *spooky* to watch and took forever to find.

As Gerry said, grids can act flaky as all get out but I've always been able to find the cause in code that "looks OK" out of context but blows up the works in context.

Dan


>I'm at a loss as to how to describe this but I'll try: I have a grid containing a couple of fields of information from a SQL table and a logical field that allows the user to check a box to select the row for inclusion in another process. Ideally, the SQL data would be read only and the column with the checkbox and caption (sparse = .f.) would be editable. The problem is that the user can select the first row using the mouse but when he tries to select another row, the data in the grid scrolls without warning, usually leaving the user at end-of-file after having checked the wrong row. It's something straight out of Whil Hentzen's "User Hostile Interface" demo. What is going on?
>
>FWIW, the method code does not use setfocus(), the recordsource is a temporary table, and in the worst case, there are up to 40 records, only 8 of which will display in the grid at one time. This seems to be similar to the problem I posted several months ago about the elevator or thumb on the vertical scroll bar moving erratically in the grid when there are more rows than will display at one time, always returning to beginning or end of file. That problem still has not been solved. Both problems may be symptoms of a major "infelicity" in the grid but if so, it seems to make it totally useless. I'm hoping I've missed something very obvious...