Re: Joins with nulls

> I guess the question is whether you want to have facts about "unknown"
> information denoted by the syntax (using NULL) or by the semantics
> (using flag values, empty strings etc.).

It's a matter of convention between the writers (or perhaps the non writers)
and the readers, regardless.

> saying Bill is age unknown is
> just as valid information as saying Bill is age 20.

It is, granted, just as valid information. But it's a different kind of information. It's not a fact about Bill's age. It's a fact about our state of knowledge about Bill's age.

Bill has an age, whether we know it or not.

If Bill's phone number is UNKNOWN, that's different. It's possible that Bill has no phone number.

Figuring out what was meant by the dog doing nothing in the night involves one more level of deduction and inference.

Likewise, figuring out what it means when data is expected, but missing, involves a level of inference (or, if you like, convention) about the real world and about the relation ship of the data in the database to it.

Many of the same issues surface when an entire row is "missing", even though no NULLS are involved.

The database knows only what it has been told to store. Including, possibly, facts that just ain't so. Received on Thu Nov 28 2002 - 13:35:44 CET