Redshift removal of explicit locks lead to missing rows duplicated rows Best resolution

Question

I have a use case where several data scientists are simultaneously using our redshift tables for tuning. We end up with missing rows and duplicate rows if they tune at the same time as importing data.

To lessen the amount of loads that would hang on those locks, they had long ago removed the explicit lock from the table.

In terms of a course of action, I was considering using kinesis, redis, or something comparable to bundle them into one import rather than inserts (not great for redshift anyways). Or if the best course of action is to deal with the hanging loads and add explicit locking back to the table.