Open jaredzh opened 2 years ago
It already is! Or at least, it's supposed to be. Could you point out where it differs?
Oh I just meant that currently we are calling a group stage function. I was thinking getting rid of that and just have map stage and reduce stage.
Oh, I see. So this would be more of a refactor. I think that's reasonable.
Make sure this is consistent with https://github.com/eecs485staff/p4-mapreduce/issues/554
Madoop should also have the "group" stage decentralized across map and reduce stages similar to EECS 485 project 4.