So for a 20 minute run, I'd develop the query on a subset (LIMIT statement?) if possible and not worry about tuning. I do a lot of data pipes and I start optimizing only if I a single run takes over an hour and I am confident that I can actually make it faster.
LIMITstatement?) if possible and not worry about tuning. I do a lot of data pipes and I start optimizing only if I a single run takes over an hour and I am confident that I can actually make it faster.