Peter,

>But if the 'only' problem with the SQL standard is the
>influence of the big vendors, why don't they get their own
>implementation more in line with the standard (or the other
>way around)?

I guess if you start with three reference implementations,
all different, by any compromise each implementation will
be wrong in some aspects.

>If I neither mis-interpret the standard nor Firebird's behaviour,
>there is a disagreement here:
>
>1. Naming constraints in DOMAIN definition
>Firebird doesn't allow the standard syntax
>CREATE DOMAIN d AS INTEGER
> CONSTRAINT d_must_be_positive CHECK (value > 0);
>
>2. Giving multiple rows in VALUE
>Firebird doesn't allow the standard syntax
>INSERT INTO t COLUMN (c1,c2) VALUES (1,2),(3,4),(5,6);

Looking over the matrix you pointed out, I notice that there
are a lot of places, small and larger, that we don't completely
support. UNION DISTINCT is one example - we don't (I think)
handle the keyword DISTINCT which is really just noise - the
default union eliminates duplicates. Easy enough to do, but
of value to whom?



Regards,

Ann
www.ibphoenix.com
We have answers.