Unless your definition of "regex" fits with whatever version of SQL your using's LIKE syntax, your going to have to read every line from the DB into memory in order to use perls regex engine. It is almost enevitable that reading from a DB will be less economical than reading a flat file.

Another problems is that, as far as I am aware, no DB will find matches that are split across record boundaries.

I personally think that unless your data has some structure to it that would allow you to make use of the DB facilities beyong storage and retreival, it doesn't make much sense to put it into a DB, but YMMV.

..and remember there are a lot of things monks are supposed to be but lazy is not one of them