Re: to_timestamp error handling.
От | Tom Lane |
---|---|
Тема | Re: to_timestamp error handling. |
Дата | |
Msg-id | 28413.1267041650@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: to_timestamp error handling. (Bruce Momjian <bruce@momjian.us>) |
Ответы |
Re: to_timestamp error handling.
|
Список | pgsql-bugs |
Bruce Momjian <bruce@momjian.us> writes: > FYI, this behavior now returns: > test=> select to_timestamp('20096010','YYYYMMDD'); > to_timestamp > ------------------------ > 2013-12-18 00:00:00-05 > (1 row) > which doesn't have the :30 but is still odd. I don't think the behavior has changed, you're merely checking it in a different timezone from the OP. The real question is whether we should throw error for out-of-range MM (or other fields). I think there are actual use cases for certain "invalid" inputs, like adding one to the day field without worrying about end of month. Perhaps there is not a use case for a month value as far out of range as this, but where would we draw the line? Anybody know what Oracle's to_timestamp does? regards, tom lane
В списке pgsql-bugs по дате отправления: