Re: [HACKERS] SET NOT NULL [NOT VALID / CONCURRENTLY]?

On 21 December 2016 at 19:01, Joel Jacobson <joel@trustly.com> wrote:

> Similar to what we (Trustly) did when we sponsored the FOR KEY LOCK
> feature to improve concurrency,
> we would be very interested in also sponsoring this feature, as it
> would mean a great lot to us.
> I don't know if this is the right forum trying to find someone/some
> company to sign up for the task,
> please let me know if I should mail to some other list. Thanks.

You'll probably get mail off list.

For what it's worth, there's a bit of a complexity here. PostgreSQL
doesn't model NOT NULL as a true CONSTRAINT. Instead it's a column
attribute. I suspect we would need to change that in order to allow a
NOT VALID NOT NULL constraint to be created.

That's at least partly why the docs say that "option NOT VALID [...]
is currently only allowed for foreign key and CHECK constraints".

Note that "[VALIDATE] acquires only a SHARE UPDATE EXCLUSIVE lock on
the table being altered" so it's already suitable for what you need.
The challenge is making it possible to create a NOT VALID constraint
for NOT NULL.

-- Craig Ringer                   http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services