bizarre AGE behaviour - Mailing list pgsql-admin

From DHS Webmaster
Subject bizarre AGE behaviour
Date
Msg-id 40461395.5324CFEE@dhs-club.com
Whole thread Raw
Responses Re: bizarre AGE behaviour  (Steve Crawford <scrawford@pinpointresearch.com>)
List pgsql-admin
Hello all,
Many thanks for this fine product.
We began encountering some unexpected date related errors this week and
after further investigation found the following. We use the postgres AGE
function in a custom function. The AGE function has begun to throw some
unanticipated results back. This has not happened before and my first
guess is that it may be leap year related. Here are some examples.

This is good...
network=# select age('04-01-04','03-01-04');
  age
-------
 1 mon
(1 row)

This isn't...
network=# select age('05-01-04','03-01-04');
         age
---------------------
 1 mon 30 days 23:00
(1 row)

Now it gets really strange......
network=# select age('06-01-04','04-01-04');
         age
---------------------
 1 mon 29 days 23:00
(1 row)

This may have been addressed in the past (I subscribe to this list) and
I just missed it, but is there a simple fix for this problem?
Thanks.
--
Bill MacArthur
Webmaster
The DHS Club, Inc.
The Best Is Yet To Come!

pgsql-admin by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: un-CLUSTER?
Next
From: Steve Crawford
Date:
Subject: Re: bizarre AGE behaviour