Re: Dawn doesn't like 1NF

From: Alfredo Novoa <anovoa_at_ncs.es>
Date: Thu, 07 Oct 2004 22:46:53 +0200
Message-ID: <9jabm09bsqod3bpj015hmft3t0mopt4fno_at_4ax.com>


On Thu, 07 Oct 2004 12:06:58 -0400, Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net> wrote:

>> The problem is that "simple" and "atomic" are not precise terms.
>>
>> For instance: Are strings simple or complex?
>>
>
>Answer by asking if a query will need to split the string. If yes, it has
>internal structure and is not atomic. If no, it is atomic. Only a single
>instance of yes renders it complex.

According to this: strings with one or zero characters are atomic and the others are not. Well, this is not too bad.

>The only way to confuse yourself over this is to mistake the internal
>patterns of a column's contents for meaningful sub-atomic values. So if
>the column "COMPANY_NAME" tends to have a lot of instances of the same
>sub-atomic values ", Inc." and ", LLC", we mistakenly conclude that the
>string must be further split to assure data correctness. But this is not
>true, the string is in fact atomic

I see an evident contradiction.

select * from customers where company_name like "%Inc."

>, it just so happens to have repeating
>patterns inside of it. So long as the strings are picked up and dropped
>onto correspondence unaltered, the string is atomic.

So now the string is atomic depending on what you do with it.

It sounds a little strange to me.

I recomend you and Laconic2 to read Date's writings in order to understand why "atomic" is not a precise term.

Regards. Received on Thu Oct 07 2004 - 22:46:53 CEST

Original text of this message