Thanks, Abdullah. I considered an approach like this, but if I were to get one error per 15 records (for example) then I'd end up having to process most records (the successful ones that were rolled-back) twice, which might end up negatively impacting performance.
↧