Re: [QUESTIONS] Join-crazy newbie may be pushing the limits :-)

Поиск
Список
Период
Сортировка
От Vadim B. Mikheev
Тема Re: [QUESTIONS] Join-crazy newbie may be pushing the limits :-)
Дата
Msg-id 34C2D131.24E714D9@sable.krasnoyarsk.su
обсуждение исходный текст
Список pgsql-hackers
Felix Morley Finch wrote:
>
> >>In article <34C1BD06.1712ABAC@sable.krasnoyarsk.su>, "Vadim B. Mikheev" <vadim@sable.krasnoyarsk.su> writes:
>
> > Felix Morley Finch wrote:
> >>
> >> I tried searching and ran into several new lessons in SQL; I ended up
> >> with this --
> >>
> >> select name from names n, keys k1, keys k2, family f1, family f2
> >> where ((k1.key ~* 'xyzzy') and (k1.id = f1.key) and (n.id = f1.name))
> >> or ((k2.key ~* 'plugh) and (k2.id = f2.key) and (n.id = f2.name));
> >>
> >> and psql never returned.  This is a test database, maybe 20 records.
> >> If I keep it down to a single keyword, it works fine.
>
> > EXPLAIN ?
> > Did you try SET geqo TO 'on=1' ?
>
> I just tried it --
>
>         WARN:Bad value for # of relations (1)
>
> I tried ON=2 which gave no WARNING but didn't fix the join.  I tried
> OFF and same result.

(So, this is problem of not just old optimizer...)

Could you post me output of

EXPLAIN VERBOSE _select_statement_ ?

Vadim

В списке pgsql-hackers по дате отправления:

Предыдущее
От: "Martin J. Laubach"
Дата:
Сообщение: Re: [HACKERS] Re: New pg_pwd patch and stuff
Следующее
От: Bruce Momjian
Дата:
Сообщение: subselects