First, I would recommend using Text::CSV_XS instead of Text::CSV. The former can handle embedded newlines and the latter cannot.

Second, you should validate the actual data that you are stuffing the database. I've written a CSV Database Validation program that you can modify for this purpose. Amongst other things, it allows you to specify unique constraints, foreign keys, and even specify a regex as a field's datatype (thus allowing much richer validation that a simple INT or VARCHAR).

Cheers,
Ovid

Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.