Re: 3 digit year problem - Mailing list pgsql-hackers

From Thomas Lockhart
Subject Re: 3 digit year problem
Date
Msg-id 3CD299A1.96C42A11@fourpalms.org
Whole thread Raw
In response to 3 digit year problem  ("Christopher Kings-Lynne" <chriskl@familyhealth.com.au>)
List pgsql-hackers
> Is there any rhyme or reason to these ISO format date parsing rules?

Yes. Though adjustments to the rules are possible, so things are not set
in concrete. There *should* be a complete description of the date/time
parsing rules in the User's Guide appendix.

> Why can't someone store the year without having to pad with zeros for years
> between 100 and 999?

To help distinguish between day numbers and years. We used to allow more
variations in the length of a year field, but have tightened it up a bit
over the years.

> What's wrong with 30-1-1 and below?  Why does 40 work and not 30?

Because "30" *could* be a day. "40" can only be something intended to be
a year. And input is not enforced to be strictly ISO-compliant, so
"30-1-1" *could* be interpreted multiple ways.
                     - Thomas


pgsql-hackers by date:

Previous
From: "Marc G. Fournier"
Date:
Subject: Re: HEADS UP: Win32/OS2/BeOS native ports
Next
From: Tom Lane
Date:
Subject: Re: HEADS UP: Win32/OS2/BeOS native ports