alt Hacker NewsIf it was less than 100 gb, he probably should have just loaded the whole thing in RAM on a single machine, and processed it all in a single shot. No S3, no network round trips, no chunking, no data warehouse.
If it was less than 100 gb, he probably should have just loaded the whole thing in RAM on a single machine, and processed it all in a single shot. No S3, no network round trips, no chunking, no data warehouse.