Shaun Thomas <sthomas(at)peak6(dot)com> writes:
> On 10/27/2011 02:13 PM, Robert Haas wrote:
>> If I had to guess, I'd bet that the second one is trying to spool the
>> resultset in memory someplace and that's driving the machine into
>> swap.
> That would be my guess too. SELECT * on a 40-million row table is a
> *lot* different than getting the count, which throws away the rows once
> it verifies they're valid.
But EXPLAIN ANALYZE throws away the rows too. There's something odd
going on there, but the information provided is insufficient to tell
what.
regards, tom lane