Markus Schaber <schabi(at)logix-tt(dot)com> writes:
> Did you try to set higher statistics targets for this columns?
Yeah, I was about to suggest the same after looking at the pg_stats
entry. It shows estimated num_distinct as 60825, ie only about 60k
distinct values in the column, which sounds kinda low (do you know
the true figure?). A larger statistics target would probably produce
a better number and hence a better join estimate.
regards, tom lane