I work for an auditing firm where we are looking for duplicate financial transactions that span approx 50M records over 21 separate tables. We are looking for matches based on, say, ABS(GROSS) and INVOICE across all 21 tables. Basically each record in each table has to be checked against the current table and each record in all the other tables. The results are complied into a separate table with a blank record between groupings. Currently this takes 72 hours. There's got to be a better way! Any suggestions?
Duty is ours, the results are God's - John Quincy Adams