From: | John Cochran <jdc(at)fiawol(dot)org> |
---|---|
To: | Oliver Elphick <olly(at)lfix(dot)co(dot)uk> |
Cc: | pgsql-patches(at)postgresql(dot)org |
Subject: | Re: timestamp patch to extend legal range of dates. |
Date: | 2003-02-05 12:01:20 |
Message-ID: | 200302051201.h15C1Keq042860@smof.fiawol.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
[snip...]
> I got my data from
> http://serendipity.magnet.ch/hermetic/cal_stud/cal_art.htm
> which also talks about proleptic calendars; that is, the current
> calendar is projected backwards (or the Julian calendar is projected
> forwards), so that there are no breaks in the sequence of dates.
>
> Does the SQL standard define what calendar it is using? My copy of Date
> & Cannan: 'SQL -- The Standard Handbook' says:
>
> "YEAR - Valid values: 0001 to 9999. This permits the specification of
> dates prior to the invention of the Gregorian calendar and assumes that
> the rules of the Gregorian calendar can be applied retrospectively."
>
> So I suggest we should use the Gregorian proleptic calendar as the
> default.
>
> Example: England Italy Proleptic
> Gregorian
>
> 15 Sep 1752 15 Sep 1752 15 Sep 1752
> 1 Sep 1752 12 Sep 1752 12 Sep 1752
> 25 Mar 1751 5 Apr 1751 5 Apr 1751
> 24 Mar 1750 4 Apr 1751 4 Apr 1751
> 6 Oct 1582 15 Oct 1582 15 Oct 1582
> 5 Oct 1582 5 Oct 1582 14 Oct 1582
> 24 Mar 1581 24 Mar 1582 3 Apr 1582
>
[snip...]
Point taken. I withdraw my patch. However, I'm going to examine date2j()
and j2date() functions a bit closer and see if I can restructure them to
eliminate the overflow problems they have. I would still like for those
functions to be capable of dealing with the full range of available numbers.
From | Date | Subject | |
---|---|---|---|
Next Message | Sumaira Ali | 2003-02-05 14:54:30 | lock.h and proc.h |
Previous Message | Sean Chittenden | 2003-02-05 02:57:07 | Re: Adding // comments (ANSI??)... |