Re: A Normalization Question
From: Jan Hidders <jan.hidders_at_REMOVETHIS.pandora.be>
Date: Wed, 14 Jul 2004 19:17:22 GMT
Message-Id: <pan.2004.07.14.19.17.50.96687_at_REMOVETHIS.pandora.be>
>
> ID Peron Color Street
> 1 brown brown brown
>
> To create an update anomaly, we need to perform an update. [...]
Date: Wed, 14 Jul 2004 19:17:22 GMT
Message-Id: <pan.2004.07.14.19.17.50.96687_at_REMOVETHIS.pandora.be>
On Wed, 14 Jul 2004 09:39:13 -0700, Neo wrote:
>> What matters there is if there are are update anomalies, and there are >> none iff there is no logical redundancy.
>
> ID Peron Color Street
> 1 brown brown brown
>
> To create an update anomaly, we need to perform an update. [...]
You already gave the example before. As already explained, it doesn't demonstrate an update anomaly in the usual meaning of the word. It only shows that a very specific update that very probably will never happen becomes easier in your representation at the cost of making a very common type of update (updating a string in a field) more difficult and costly.
- Jan Hidders