Re: object algebra

From: Jonathan Leffler <jleffler_at_earthlink.net>
Date: Sun, 07 Mar 2004 05:14:57 GMT
Message-ID: <ldy2c.22867$yZ1.3107@newsread2.news.pas.earthlink.net>

Neo wrote:

 >Jonathan Leffler wrote:
>>Please can you explain the definition of a relation which you think 
>>means that the RDM must create nulls?  And can you give a reference to 
>>where the version of the RDM you are using or assuming is defined?  It 
>>doesn't seem to be the same as the version(s) I'm used to.

>
> I am using the definition which matches C. J. Date's definition in his
> book "An Introduction to Database Systems", 6th Ed, ISBN
> 0-201-54329-X, in Chapter 4, section 3, under the heading of Relation.
> A relation consist of two parts, a heading and a body. The heading
> consists of a fixed set of attributes. The body consists of a set of
> tuples. Each tuple consists a set of values corresponding to the
> heading.

And where in that chapter are nulls described? I dug out my copy of the book (it was in peaceful retirement since I have the 7th and 8th editions for current references), and an admittedly quick scan did not show me even a footnote mentioning nulls, nor any mention of missing information - which is where nulls appear. Indeed, I would not expect to find any discussion of nulls in that section of the book until chapter 8, which is on SQL.

>> Your previous answer contained no definition of a relation which >> explains why nulls are an integral part of the RDM.

> The defintion of a relation and how NULLs are incurred was to be
> inferred from the analogy about Rover callecting data on Mars.

A rather less than compelling example, IMO - and the main inference I drew was that you are not very good at explaining what you're talking about.

> The operation (INSERT) was to be inferred from the analogy. A NULL
> occurs when reality can't provide a value required by a tuple.

This is not compelling. If you can't provide a value for a column, then you can't insert the row - in a relational DBMS (as distinct from, or opposed to, an SQL DBMS).

> Please see other postings for different
> views on this topic.

No, I'm about to put this thread on my 'ignored' list.

-- 
Jonathan Leffler                   #include <disclaimer.h>
Email: jleffler_at_earthlink.net, jleffler_at_us.ibm.com
Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/

Received on Sat Mar 06 2004 - 23:14:57 CST