Re: Does Codd's view of a relational database differ from that ofDate&Darwin?[M.Gittens]

From: Alexandr Savinov <>
Date: Tue, 07 Jun 2005 19:25:03 +0200
Message-ID: <42a5d881$>

erk schrieb:
>>The problem is
>>if we want to keep meaningless columns in one table or optimize
>>representation. I mean that independent of our desire or schema any
>>object in the database will formally have all dimensions. The question
>>is only how we *represent* the database semantics. In one case null is
>>written explicitly. In other cases it still exists as a value of an
>>object along meaningless dimensions but is not explicitly written
>>because of good schema. For database and for the model this object still
>>has null as a value along this dimensions independent of how it is

> So null is there even if it's not? Then I much prefer my nulls
> not-there.

Assume you have two tables Products =<Price, Size> and Personel = <Age, Salary>. Then you have a record in Products r=<$100, big>.

Question: What Age has the record r?

Answer: null.

Comment: Even if the null value is not stored explicitly because of such two-table design, it does not change the data semantics, i.e., record r in the canonical semantics has no value for dimension Age or it does not exist in this dimension. If you transform equivalently the model then you have a change to see them explicitly as real null values in 4-columns table. Independent of how you represent your data and how you decompose schema the semantics reamains the same.

Received on Tue Jun 07 2005 - 19:25:03 CEST

Original text of this message