Re: Plan for update ... where a is not distinct from b

On 2017-11-28 09:35:33 -0500, Tom Lane wrote:
> "Peter J. Holzer" <hjp-pgsql@hjp.at> writes:
> > I noticed that an update was taking a long time and found this:
> > [ crappy plan for join on IS NOT DISTINCT ]
>
> Yeah, there's no optimization smarts at all for IS [NOT] DISTINCT.
> It can't be converted into a merge qual, nor a hash qual, nor an
> indexscan qual.

Pity. I expected IS NOT DISTINCT to be treated pretty much like =, given
that it is just a more naive equality test. In particular, since
PostgreSQL stores NULL values in indexes (unlike Oracle) I expected it
to be able to use an index scan.

> In principle this could be improved, but given how much work it'd be
> and how seldom anyone complains, it's not likely to happen anytime soon.

Yeah, IS [NOT] DISTINCT is pretty obscure. I guess not many people use
it.
       hp

--   _  | Peter J. Holzer    | we build much bigger, better disasters now
|_|_) |                    | because we have much more sophisticated
| |   | hjp@hjp.at         | management tools.
__/   | http://www.hjp.at/ | -- Ross Anderson <https://www.edge.org/>