HomeLog InSource


[View] [Short] [Hash] [Raw]
Okay, so our "unsorted" sorted intersection algorithm is even more complex than we thought.

Assuming our indexes are too big to fit in memory, it's a fair claim that the result will be too. So we need to produce our output direct-to-file. That's even harder than it sounds because our files are append-only and we want to keep it in order.