Thread: BUG #15389: Fill zero in milliseconds of a timestamp
The following bug has been logged on the website: Bug reference: 15389 Logged by: Luca Taddeo Email address: taddeoluca96@gmail.com PostgreSQL version: 9.6.10 Operating system: Windows 7 Service Pack 2 Description: if I try to convert a string "2014-09-26 17:17:45.990" into timestamp become "2014-09-26 17:17:45.99" I don't think is correct because if I try to convert in Java become 099.
Hello > I don't think is correct because if I try to convert in Java become 099. "2014-09-26 17:17:45.099"? I think Java is incorrect here. Usually trailing zeros can be omitted. Also postgresql timestamp has resolution 1 microsecond, so with trailing zeros it must be "2014-09-26 17:17:45.990000", not"2014-09-26 17:17:45.990". regards, Sergei
On Tuesday, September 18, 2018, Sergei Kornilov <sk@zsrv.org> wrote:
Hello
> I don't think is correct because if I try to convert in Java become 099.
"2014-09-26 17:17:45.099"?
I think Java is incorrect here. Usually trailing zeros can be omitted.
Also postgresql timestamp has resolution 1 microsecond, so with trailing zeros it must be "2014-09-26 17:17:45.990000", not "2014-09-26 17:17:45.990".
Except the decimal in that construct is a simple part separator and not a fractional indicator. You cannot drop trailing zeros in the last part in the same way you cannot drop trailing zeros when describing one hundred (100) because if you do you get the number one (1) instead.
David J.
.
>>>>> "David" == David G Johnston <david.g.johnston@gmail.com> writes: >> I think Java is incorrect here. Usually trailing zeros can be >> omitted. Also postgresql timestamp has resolution 1 microsecond, so >> with trailing zeros it must be "2014-09-26 17:17:45.990000", not >> "2014-09-26 17:17:45.990". David> Except the decimal in that construct is a simple part separator David> and not a fractional indicator. Both the SQL spec and ISO 8601 beg to differ. -- Andrew (irc:RhodiumToad)