SELECT INTO TABLE busted?

Поиск
Список
Период
Сортировка
От Tom Lane
Тема SELECT INTO TABLE busted?
Дата
Msg-id 17061.918190417@sss.pgh.pa.us
обсуждение исходный текст
Ответы Re: SELECT INTO TABLE busted?  (Bruce Momjian <maillist@candle.pha.pa.us>)
Re: SELECT INTO TABLE busted?  (Bruce Momjian <maillist@candle.pha.pa.us>)
Список pgsql-hackers
I've been seeing the following behavior with the CVS sources for the
last several days:

$ psql regression
regression=> SELECT * INTO TABLE ramp   FROM road  ;
SELECT
regression=> \d ramp
Couldn't find table ramp!
regression=> SELECT * INTO TABLE ramp   FROM road  ;
ERROR:  ramp relation already exists
regression=> SELECT * FROM ramp ;
< works fine, produces plenty of output... >

If you exit psql and start a new session, "ramp" has disappeared
entirely.  This is causing the create_view regression test to fail,
since it depends on a table made like this in a prior test.

My guess is that this is an unwanted side effect of the "temp table"
code.  Is anyone else seeing the same, or am I looking for a
platform-specific bug?  (gcc 2.7.2.2 on HPUX 9.03, for the record.)
        regards, tom lane


В списке pgsql-hackers по дате отправления:

Предыдущее
От: Bruce Momjian
Дата:
Сообщение: Re: [HACKERS] Adding some const keywords to external interfaces
Следующее
От: Bruce Momjian
Дата:
Сообщение: Re: SELECT INTO TABLE busted?