RE: oracle timestamp internal storage and precision

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Wed, 19 Dec 2012 03:21:37 -0500
Message-ID: <00e301cdddc1$de0d1b80$9a275280$_at_rsiz.com>



In addition to the fine replies from Hans and Jared, I blogged including some stuff about nulls and vsize.

http://rsiz.wordpress.com/2012/12/19/some-stuff-about-dates-and-timestamps-i n-oracle/

After all y'all read this, please chime in individually to Oracle for enhancement requests to add rsize (or rsiz (!)) as an alternative to vsize showing the storage size of null values, and allowing .FF [1..9] as a legal part of the format value for a date to make consistent date/time formatting easier (even though they'll always be zeroes for date, this would add the possibility of consistency) as you see fit.

While an rsiz function could be written as an add-on, not having it as a built-in would make it not worth the effort, so really only Oracle can do this.

mwf

-----Original Message-----

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Walker, Jed S
Sent: Tuesday, December 18, 2012 7:50 PM To: oracle-l
Subject: oracle timestamp internal storage and precision

I could swear that a while back I read or saw a presentation from someone where they discussed how Oracle stores the internal timestamp. They said it was stored as a number and that in order to not expire to early (like UNIX time) the storage would, over time, sacrifice sub-second precision to extend the date range. I tried searching for this, but have found nothing. I wonder if I'm remembering wrong, or if maybe it just isn't something that has been published.
Do any of the Oracle internal experts out there know?

--

http://www.freelists.org/webpage/oracle-l

--

http://www.freelists.org/webpage/oracle-l Received on Wed Dec 19 2012 - 09:21:37 CET

Original text of this message