J
James Palmer
Guest
In my opinion this is quite a dangerous inconsistency. I've raised it before, and in fact I think I've raised an enhancement request. In my previous job we had quite a few cases where we used find by rowid to refind a record. The person who wrote the logic in the first place had read about -rereadnolock, but not about this limitation. As a result there were a number of reported instances where the logic resulted in old data being written back to the database when 2 processes updated the same record in succession. I've found the enhancement: community.progress.com/.../-rereadnolock_to_be_honoured_by_rowid_finds
Continue reading...
Continue reading...