Incremental seeding possible (split large csv into smaller ones)?

[jpwynn]

As another issue mentioned, the memory footprint is pretty ginormous.

A 80K row csv that is 16MB in size (zipcodes) gives an out of memory error on Heroku (512MB) when attempting to run rake db:seed

The ability to either (a) not load the whole shebang into memory at once, or (b) chunk the seed data into multiple files and seed them one at time would be handy.