I have business case where a listener would read a set of 50 files and create work objects simultaneously. We have a logic to identify duplicates in the process flow where an RDB-List identifies the duplicate Work object list. But when work objects are created within a time gap of few micro seconds, e.g. W-1 and W-2 where W2 is a duplicate of W1, the RDB-List doesn't return W2 as duplicate W1 as W1 is still not committed into DB when W2 duplicate logic is being executed.

Any suggestions on how to handle this scenario?