At 01:00 AM 10-04-01 +0000, you wrote:

>--- In ib-support@y..., Helen Borrie <helebor@d...> wrote:
> > At 11:08 AM 09-04-01 +0200, you wrote:
> > >Helen,
> > >
> > > > >If I define a computed column like the one in documents:
> > > > >full_name computed by (first_name || ' ' || last_name)
> > > > >It works only if non of the two composing columns is null.
>Otherwise, the
> > > > >value is null if any of the two is null.
> > > >
> > > > Correct. Value + NULL evaluates to false.
> > >
> > >As a matter of fact, Value + NULL evaluates to NULL.
> > >I guess you meant that.
> >
> > No, I meant what I said.
>
>Then I don't understand your reply.
>
>I thought that the original message said "Value + NULL is NULL". You
>said "Correct", which I (in my admittedly very simple [and I strive to
>keep it that way] mind) took as agreement that "Value + NULL is
>NULL/unknown". However, you then continued to say "Value + NULL
>evaluates to FALSE".
>
>Either Value + NULL evaluates to NULL/unknown, or Value + NULL
>evaluates to FALSE ... even in the SQL standard, you can't have it
>both ways.

Au contraire (being academic here, for which reason I didn't bother to respond). I made a statement and somebody assumed I wrote something I didn't meant to. I meant what I wrote. You can't add NULL to anything because NULL is not a value, i.e. Value + NULL = NULL is absurd in Boolean terms. We train databases to recognise absurd propositions as "false" (even though "false" and "absurd" are not the same thing) and return NULL to them instead of returning a "false proposition" error.

The proposition Value + Value evaluates to true because it is possible to derive a result from the arguments. As a result of its evaluating to true, the arithmetic is performed and a result is returned.


Cheers,
H.
All for Open and Open for All
InterBase Developer Initiative · http://www.interbase2000.org
_______________________________________________________