Обсуждение: check contraints incorrectly reject "null"

Given a table definition like:

create table foo (i integer check (i > 0));

I noticed the following works in Oracle but fails in Postgres:

insert into foo values(null);

I was curious about what the standard might say, and had been
meaning to buy Date's book for some time, so broke down and
did so.

According to Date, a check contraint should fail if the expression
evaluates to false.  It appears that Postgres only passes the
check constraint if it evaluates to true.  In three-valued logic,
these statements aren't equivalent.  He has a paragraph about
nulls and check contraints in chapter 14, I believe, and his
explanation makes it clear that Oracle is right, Postgres wrong.

It's easy to fix by adding a check for null to the constraint,
and afterwards the SQL still works with Oracle, but it's still
a bug...



- Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert
Serviceand other goodies at http://donb.photo.net.
 

Don Baccus <dhogaza@pacifier.com> writes:
> According to Date, a check contraint should fail if the expression
> evaluates to false.

And SQL92 says:
        A table check constraint is satisfied if and only if the specified        <search condition> is not false for
anyrow of a table.                              ^^^^^^^^^
 

so they agree: a constraint that yields NULL should be considered
to pass.  A tad nonintuitive, but who am I to argue...

I have fixed several bugs recently having to do with incorrect
evaluation of three-state boolean logic.  I'll take care of this one.
        regards, tom lane