Well then perhaps I'm off base but if he's trying to keep out duplicates can't some type of query be written to ignore duplicates based on the unique id ?