I'd look into what bbfu says above first about an import facility for your database, and if there isn't any import facility, or if there's some filtering/munging that the import facility just doesn't provide that you want to do in perl (then again you can filter/munge and output another CSV file to import, which if its a large dataset, can be a faster way to do it), then you could look into the AnyData module (or another CSV module), and the insert_hash example from the latest DBI perldocs.