Re: The BOOLEAN data type

From: Paul Vernon <paul.vernon_at_ukk.ibmm.comm>
Date: Tue, 1 Apr 2003 11:38:20 +0100
Message-ID: <b6c0ed$1inq$>

"--CELKO--" <> wrote in message
> [...] Booleans are
> usually, but not always, a sign of bad programming.

that would surprise me if it were true, but then what has programming got to do with database theory?

> Someone is
> storing the state of the database at one point in time as a flag.

I'm not at all sure what that means

> It
> is a computed column and we all know better than to store redundant
> data like that.

Storing (uncontrolled) redundant data is indeed bad, but I can't see that Boolean valued attributes are more likely to be storing redundant data than say integer valued attributes.

On the other hand, I do agree that it is often better to use the presence or absence of a row to represent some business boolean value rather than an boolean valued attribute. Rather like attributes called 'status', there is often more to capture about the change in value of a boolean than just the value change itself, not least things such as the time the change occurred, who made it and why.

Paul Vernon
Business Intelligence, IBM Global Services Received on Tue Apr 01 2003 - 12:38:20 CEST

Original text of this message