Michael Fuhr <[EMAIL PROTECTED]> writes:
> That's a lot of data -- are you aware that psql (via libpq) fetches
> the entire result set before displaying it? In most cases 18444
> rows wouldn't be a problem, but with rows that wide it becomes a
> big problem because the client has to store it all in memory. I
> wonder if that's causing psql to segfault, although I'd expect a
> graceful error like "out of memory for query result" unless maybe
> psql consumes so much memory that the OS has problems.
I'm wondering the same --- psql is definitely designed to survive
out-of-memory:
regression=# select * from tenk1 a, tenk1 b limit 2000000;
-- time passes ...
out of memory for query result
regression=#
Emilia might have found some corner case where it doesn't, though;
perhaps a malloc call that's not error-checked. A stack trace from
the psql core dump would be useful.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
