that's exactly my point I have traced it to that point. I put a break point on both the REPLACE and the SCAN and checked the contents of the view on each side. Those are the commands where the data is disappearing. I know it doesn't make sense but that is what I am seeing.

Because it works the first time through I feel it has to be something to do with the view.

I am experiencing one other problem with my remote views. For some reason they periodically wont TABLEUPDATE. If I issue the same TABLEUPDATE subsequent to this it works.

I have a funny feeling it is the update that is somehow corrupting the view. If I close and reopen the view at the beginning of each pass I have no problem.

I have tried deleting the view and recreating it to no avail. I will try delete view/pack database/validate database/add view and retest. Also another database if that does not work.

Regards
Geoff Scott


>>No, I have checked that. It was one of the reasons I went to SCAN. When I open the browse window in debug at the SCAN command the data is there. The I step into the SCAN (ie execute the SCAN command) and then switch back to the browse window and the data is gone. It should be on the first record, which it is (sort of) but all the fields are empty. Well at least it steps into the SCAN code block which would indicate there is a record there.
>>
>>I did notice that in later processing where I first detected the problem there are no records in the cursor as opposed to being empty.
>
>Well there is nothing about either the REPLACE ... ALL or the SCAN/ENDSCAN that would delete records so there has to be other code that is doing that to you. I would next try SET STEP ON and step through the code in the debugger to see when the record(s) is being deleted.