Re: Normalization and Derived Information
From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Thu, 23 Sep 2004 10:36:15 -0400
Message-ID: <v0nuic.rc9.ln_at_mercury.downsfam.net>
>
> The first thing I want to say is that data normalization, considered
> rationally, is a design principle that you can use to avoid certain
> anomalies. However, too many people, including some of the denizens of
> this newsgroup, treat it as the sole inerrant sign of allegience to the
> One True
> Faith (OTF) of the Relational Data Model (RDM). Any design, like star
> schema, that deliberately settles for anything less than FULL
> NORMALIZATION (shouting implied), is excommunicate and anathema and it
> identifies the perpetrator as an infidel!
>
> This is nonsense. It amounts to turning a useful concept into mumbo
> jumbo.
>
> I think, but I'm not sure, that the Extended Price is redundant with the
> quantity and unit price in the same row. So it really is one fact in more
> than one place, and probably violates some normalization rule. Now ask
> me
> if I care. The extra few bytes of storage are probably worth it.
>
> Just for grins, most of the usual star schema designs for this kind of
> data has ONLY the extended price in the fact table.
> The advantage of the extended price is that it can be summed. (other
> trivia about star schema design omitted).
Date: Thu, 23 Sep 2004 10:36:15 -0400
Message-ID: <v0nuic.rc9.ln_at_mercury.downsfam.net>
Laconic2 wrote:
>
> "Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message
> news:367sic.063.ln_at_mercury.downsfam.net...
>> I'm wondering about the precise status of derived columns in terms of >> 1nf, 2nf and so forth.
>
> The first thing I want to say is that data normalization, considered
> rationally, is a design principle that you can use to avoid certain
> anomalies. However, too many people, including some of the denizens of
> this newsgroup, treat it as the sole inerrant sign of allegience to the
> One True
> Faith (OTF) of the Relational Data Model (RDM). Any design, like star
> schema, that deliberately settles for anything less than FULL
> NORMALIZATION (shouting implied), is excommunicate and anathema and it
> identifies the perpetrator as an infidel!
>
> This is nonsense. It amounts to turning a useful concept into mumbo
> jumbo.
>
> I think, but I'm not sure, that the Extended Price is redundant with the
> quantity and unit price in the same row. So it really is one fact in more
> than one place, and probably violates some normalization rule. Now ask
> me
> if I care. The extra few bytes of storage are probably worth it.
>
> Just for grins, most of the usual star schema designs for this kind of
> data has ONLY the extended price in the fact table.
> The advantage of the extended price is that it can be summed. (other
> trivia about star schema design omitted).
The only thing I could possibly add to this excellent post is this: "the computer is made for man, man is not made for the computer."
In the OP I did not actually ask if I *should* do this, since my mind was made up on that ten years ago. I just wanted to know precisely what I would be charged with in court.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Thu Sep 23 2004 - 16:36:15 CEST