Stephen no one is saying its equal to anything, I look at NULL as undefined
values, so when you are testing a value against NULL, you are merely testing
to see if it was defined, as if I wanted to get all values that are not
defined, I would test for =NULL.  As I mentioned before you can set a field
to "No value known" or NULL, but you can't query on it, is it just me that's
having a hard time comprehending this.  I've worked with this for many
years, though it for some reason never bothered me until now, as I was
trying to write something where this came into a big play.

Ilya

-----Original Message-----
From: Stephen Clouse
To: Bart Lateur
Cc: [EMAIL PROTECTED]
Sent: 10/23/01 4:10 PM
Subject: Re: = NULL vs. IS NULL

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, Oct 23, 2001 at 11:56:39PM +0200, Bart Lateur wrote:
> My personal opinion is to disagree. To me, NULL means "empty". It is
not
> the same as a zero length string.

When did I say it was?  In fact, I thought I explicitly stated that
Oracle was 
wrong in this regard.  A zero-length string is very different from the
concept 
of NULL, which is "no value known".  Zero-length string, OTOH, is a
known value, 
one with no characters.

>                                   But empty is empty, thus NULL=NULL.
> 
> Nitpicking that NULL != NULL, is only making our life harder.
>
> Last week, there was a similar discussion going on, on the Perl6
mailing
> lists, with regards to NaN (Not A Number). Is NaN==Nan, or NaN!=NaN?

ANSI is correct in the case of NULL.  If something has no value, how can
you 
equivocably say that it's equal to anything, even something else with no
value?  
And if you say the value is empty then the value isn't NULL because you
*know* 
the value; it's a zero-length string.  So the whole argument becomes
moot 
anyway.

Of course, if you're using Oracle, '' gets treated as NULL, but that's
their 
fault.  (I'd wager this whole thread stems from an Oracle user to start
with :)  
Perhaps they'll finally retire that broken behavior in version 10.  Of
course, 
they said they were going to do it in version 8....

- -- 
Stephen Clouse <[EMAIL PROTECTED]>
Senior Programmer, IQ Coordinator Project Lead
The IQ Group, Inc. <http://www.theiqgroup.com/>

-----BEGIN PGP SIGNATURE-----
Version: PGP 6.5.8

iQA/AwUBO9XquwOGqGs0PadnEQLNWwCgi16nojePo2E4tHzlKlVi1whTrioAoKzT
W0P+kERK+EyKll34RyLXSwb3
=lven
-----END PGP SIGNATURE-----

Reply via email to