Re: Slow subquery on large dataset

bob_bamber(at)hotmail(dot)com (Bob) writes:
> I'm having some performance issues when querying a couple of tables
> containing a large amount of data.

> foo=# SELECT data_id FROM capacity_data WHERE data_id NOT IN (SELECT
> data_id FROM capacities);

Try PG 7.4, or if you're on 7.4, try increasing sort_mem. IN and NOT IN
have pretty sucky performance in earlier releases. (If you can't
upgrade for some reason, you could try using EXISTS instead of IN.)

regards, tom lane