As the maintainer of DBM::Deep, I need to echo tilly here. While dbm-deep can allow you to address such large datasets (>4G requires 64-bit Perl just because 32 bits only addresses 4G and change), it is going to be very slow. Just building the 4G file can take a very long time. Sorting a large array (>10k entries, roughly) in dbm-deep will take a very very long time.
in reply to Re: Memory Efficient Alternatives to Hash of Array
in thread Memory Efficient Alternatives to Hash of Array
While one of dbm-deep's use cases is dealing with data that's too large to fit in RAM, you really want to have that happen with lookups, not when dealing with large datasets. There are other languages and setups better suited for this, such as Erlang or CouchDB.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?