Re: BUG?: timestamp without TZ created as timestamp *with* TZ

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: BUG?: timestamp without TZ created as timestamp *with* TZ
Дата
Msg-id 25828.1030629404@sss.pgh.pa.us
обсуждение исходный текст
Ответ на BUG?: timestamp without TZ created as timestamp *with* TZ  (Jean-Christian Imbeault <jc@mega-bucks.co.jp>)
Ответы Re: BUG?: timestamp without TZ created as timestamp  (Thomas O'Dowd <tom@nooper.com>)
Список pgsql-general
Jean-Christian Imbeault <jc@mega-bucks.co.jp> writes:
> It seems that even though I declare a columns to be of type timestamp
> without time zone it gets created as a data type *with* a time zone?

That is the behavior in 7.2; it's a transient state to help people
migrate from our old not-very-SQL-compliant datatype names.

In 7.3 "timestamp" will mean "timestamp without time zone" per spec.

> Here my SQL:
> create table MEMBERS (
> last_login              timestamp without time zone [0],
> );

> I then do a "\d members" and get:

>   last_login      | timestamp with time zone[] |

There's also a nasty little bug in the 7.2 system catalogs: the
array-of-timestamp-without-tz datatype is mistakenly linked to the
timestamp-with-tz element type.  You can fix this with a quick UPDATE to
pg_type; in the source code the correction looks like

403c403
< DATA(insert OID = 1115 ( _timestamp  PGUID    -1 -1 f b t \054 0      1184 array_in array_out array_in array_out d x
_null_)); 
---
> DATA(insert OID = 1115 ( _timestamp  PGUID    -1 -1 f b t \054 0      1114 array_in array_out array_in array_out d x
_null_)); 

The actual SQL command to give is in the archives somewhere, if you
need more help.

            regards, tom lane

В списке pgsql-general по дате отправления:

Предыдущее
От: "Jules Alberts"
Дата:
Сообщение: Re: Securing sensitive information
Следующее
От: Stephan Szabo
Дата:
Сообщение: Re: Deleting foreign key constraints