Re: What does this NULL mean?

From: David Portas <REMOVE_BEFORE_REPLYING_dportas_at_acm.org>
Date: 11 Dec 2005 05:48:44 -0800
Message-ID: <1134308924.277933.121900_at_o13g2000cwo.googlegroups.com>


John Eggert wrote:

>
> Ok, I've been somewhat ignoring this entire topic lately, but I have a real
> world situation that I'd like to see a solution for. I am currently moving
> the operator reporting for an industrial process from a paper system to a
> computer based system. One thing that is occuring with the paper system is:
>
> An operator does what are called 'rounds'. They do certain things at a
> certain time. Most of these are manual checks of various circuit
> parameters. Take one of these as an example: The amount of activated carbon
> per tank ( I like this one because there is no automated method for this
> measurement at this time ). They then write this number down. On occasion,
> it is not possible to get the value. They are usually busy somewhere else.
> This information (the fact that the operator missed a round) is indicated
> by the blank cell in the hand written report sheet. I would interpret this
> as a NULL value. How can I capture the same data (the missing information
> IS the data) in a database? Just to be clear, the empty cell in the
> completed operator's report sheet IS USEFULL DATA. It isn't 0, it isn't
> unknown, it is missing data. The fact that >>the data is missing<< is the
> data I need to capture. I need to be able to quantify how frequently the
> data is missing, if it is any particular operator or any particular part of
> the process etc. I am using "NULL" as a default value. Given the long
> debate on the good bad and ugly of NULL, what is wrong with that? Isn't
> this exactly the situation where a NULL value is needed, appropriate and
> functionally usefull?
>
> With respect to the normalization aspect. How can one better normalize a
> system so that one can capture the fact that someone did not fill in
> something that should normally have been filled in but for unknown reasons
> did ( and possibly could ) not? If I have single attribute tables I can
> still have the missing data. How can I further normalize a single number?
>
> Cheers.
>
> JE

Have you read:

http://web.onetel.com/~hughdarwen/TheThirdManifesto/Missing-info-without-nulls.pdf

-- 
David Portas 
SQL Server MVP 
--
Received on Sun Dec 11 2005 - 14:48:44 CET

Original text of this message