Re: Slow planing...

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: Slow planing...
Дата
Msg-id 4902.1575643524@sss.pgh.pa.us
обсуждение исходный текст
Ответ на Re: Slow planing...  (Mladen Marinović <mladen.marinovic@kset.org>)
Список pgsql-general
=?UTF-8?Q?Mladen_Marinovi=C4=87?= <mladen.marinovic@kset.org> writes:
> After a couple of hours of trying different stuff, set enable_mergejoin  =
> off made the planning time look better: Planning time: 0.322 ms
> Any ideas why this helps?

Hmm, possibly because it's reducing the number of calls to
get_actual_variable_range.  Do you have an indexed column,
with a lot of recently-dead entries at the end of its range,
that is a join key in the slow query?  If so, you might be
hitting the problem that was fixed (not for 9.6) here:

https://git.postgresql.org/gitweb/?p=postgresql.git&a=commitdiff&h=3ca930fc3

            regards, tom lane



В списке pgsql-general по дате отправления:

Предыдущее
От: Mladen Marinović
Дата:
Сообщение: Re: Slow planing...
Следующее
От: Adrian Klaver
Дата:
Сообщение: Re: Insert Table from Execute String Query