Alvaro Herrera <alvherre@xxxxxxxxxxxxxxxxx> writes: > Hmm, I'm placing bets on the bpchar weirdness. I'd try getting rid of > that and using plain varchar for all the columns. That's certainly what's inhibiting it from considering an indexscan on the larger table. I'm not as convinced as the OP that a nestloop indexscan is really going to win compared to the hash plan, but if the comparison value is varchar then an index on a bpchar column is simply not useful --- at least not unless you stick an explicit cast into the query, so that the comparison will have bpchar rather than varchar semantics. regards, tom lane -- Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
