On Fri, Feb 28, 2014 at 5:01 PM, Damon Snyder <damon@huddler-inc.com> wrote:
> The primary query that I'm trying to optimize executes in about 1600ms on my
> laptop and about 800ms on production-like hardware (more for the score
> version). My target is to get the data fetch down below 100ms if possible.
Could you post some explain analyze of those particular queries?
> If you have any suggestions it would be greatly appreciated. Am I missing
> something obvious? Is there a logically equivalent alternative that would be
> more efficient?
I'd suggest de-normalizing a bit. For instance, why don't you put the
score right into the object? I'm sure the indirection is hurting.