Unless the data needs to be human-readable, why not just use
DBM::Deep? That allows you to treat your serialization directly as a Perl data structure. And, for bulk-additions, just whip together a little script that adds the stuff to the Perl data structure that is your file.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?