Re: BUG #16472: Bug in to_timestamp ? - Mailing list pgsql-bugs

From 鄭旭宏
Subject Re: BUG #16472: Bug in to_timestamp ?
Date
Msg-id CAJsb5a0mUq+Q0DNeMuoNmgSLOOxwq9=KYKc471KQoQP8JVUWGA@mail.gmail.com
Whole thread Raw
In response to Re: BUG #16472: Bug in to_timestamp ?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-bugs
timezone = Asia/Taipei  
version = PostgreSQL 9.6rc1, compiled by Visual C++ build 1800, 64-bit

Tom Lane <tgl@sss.pgh.pa.us> 於 2020年6月2日 週二 上午12:18寫道:
Alexander Korotkov <a.korotkov@postgrespro.ru> writes:
> On Mon, Jun 1, 2020 at 1:38 PM PG Bug reporting form
> <noreply@postgresql.org> wrote:
>> I execute a update SQL as below :
>> update Table1 set date1=to_timestamp('19790701000000', 'YYYYMMDDHH24MISS')
>> where .....
>> date1 is a timestamp type column, the result in DB is  1979-07-01 01:00:00
>> where come from that '01' hour ?

> It's likely related to clock shift in your timezone.

July 1 would be an odd time for a seasonal DST shift ... but quickly
thumbing through the tzdata files, I see one for Asia/Taipei on that date
in 1979.  So I guess that the OP is using that zone, and the answer to the
question is "midnight of that date did not exist in Taiwan; their clocks
shifted forward from 23:59:59 to 01:00:00".

                        regards, tom lane

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: pgbench bug / limitation
Next
From: Michael Paquier
Date:
Subject: Re: