> btw, it appears that SQL99 (haven't checked SQL92) specifies that
>
> test=# select (1,2,3) = (1,2,null);
> ?column?
> ----------
>
> (1 row)
>
> should return FALSE, not NULL. I haven't looked to see if this is
> consistant with other comparisons involving NULL.
This is interesting. If "(1,2,3) = (1,2,null)" is legal from the point
of view of the standard, isn't "(1,2,null) = (1,2,null)" legal
comparison too? no?
BTW, I would like to object the "null = null" support in
PostgreSQL. It is against the standard. That's enough reason to reject
the syntax.
--
Tatsuo Ishii