On 20/12/06, Paul Vinkenoog <paul@...> wrote:

> Nick, to get back to your original posting:
>
> > when I do
> > select * from mytable where test = 0;
> > I would expect zero rows and when I do
> > select * from mytable where test is null;
> > I would expect 1 row
> >
> > yet it's the other way round
>
> That (and not the value you see) is proof that the field is not NULL,
> but really 0.

that's why I did the test, in case I was being shown something
different from the actual data, what I don't get is why. AFAIK it
should be null

> The rotten thing is the no-default case, where people *see* a 0 when
> they select but the field is really null. Firebird allows this
> situation to exist, also allows such a database to be backed up, but
> refuses to restore it (except with -n).