Re: relations aren't types?

From: Lauri Pietarinen <lauri.pietarinen_at_atbusiness.com>
Date: Tue, 30 Dec 2003 01:57:57 +0200
Message-ID: <bsqf7f$303$1_at_nyytiset.pp.htv.fi>


Costin Cozianu wrote:

>

> The elementary test of "atomicity" is whether you can decompose a
> value of that type into its components.

My understanding is that in the scope of relational databases, atomicity is defined in terms of whether the relational operators can "see" the value
or not without the help of "non-relational", or scalar operators.

So, in e.g.

session(session_id, session_timestamp) the timestamp will be atomic

and in

session(session_id, session_date, session_time, session_msec) the date, time and millisecond parts will each be atomic, respectively.

In

person(person_id, person_name, person_image) person_image will be atomic as far as the RDBMS is concerned, no matter how complex the inner structure of the image is and in

person(person_id, person_name, image_id) image(image_id, image_date, etc...)
pixel(image_id, pixel_id, pixel_color)

the image is not atomic, because it's components can be accessed via relational operators.

A bit like protons, neutrons and electrons being the "atoms" of chemistry even though these particles can be further decomposed into subparticles, but not with the tools available to chemists.

best regards,
Lauri Pietarinen Received on Tue Dec 30 2003 - 00:57:57 CET

Original text of this message