From: "Josh Berkus" <josh@agliodbs.com>
Subject: Re: [SQL] Bug with Daylight Savings Time & Interval 
To: Tom Lane <tgl@sss.pgh.pa.us>, josh@agliodbs.com
Cc: pgsql-sql@postgresql.org, pgsql-bugs@postgresql.org
Date: Mon, 20 May 2002 22:57:46 -0700
Message-ID: <web-1465902@davinci.ethosmedia.com>
In-Reply-To: <17307.1021949260@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit

Tom and Tom,

> This isn't a bug per the existing definition of INTERVAL.  '250 days'
> is
> defined as '250*24 hours', exactly, no more no less.  When you move
> across a DST boundary you get behavior like the above.
 
> I've opined several times that interval should account for three
> separate units: months, days, and seconds.  But our time-meister
> Tom Lockhart doesn't seem to have taken any interest in the idea.

I beg to differ with Tom L.  Even if there were justification for the
addition of an hour to a calculation involving only days, which there
is not, there are two bugs with the existing behavior:

1. You do not lose an hour with the end of DST, you just gain one with
the beginning of it (until you wraparound a whole year, which is really
confusing), which is inconsistent;

2. Even if you justify gaining or losing an hour through DST in a
'+days' operation, changing the TIMEZONE is a bizarre and confusing way
to do it.  I don't fly to Colorado on April 7th!

While this needs to be fixed eventually, I need a quick workaround; is
there a way to "turn off" DST behavior in PostgreSQL?

Further, it seems that the whole "Interval" section of Postgres,
possibly one of our greatest strengths as a database, has languished in
the realm of inconsistent behavior due to lack of interest.  Is there
anything I can do without learning C?  

-Josh Berkus