Re: pgsql: CREATE INDEX ... INCLUDING (column[, ...])

Peter Geoghegan-3 wrote
> On Fri, Apr 8, 2016 at 11:31 AM, Robert Haas <

> robertmhaas@

> > wrote:
>>> I assume that the problem here is larger than just failure to adhere to
>>> C89 comment style. Was this patch really ready to commit? I'm not very
>>> happy that such a large patch went from "Needs review" to "Committed" in
>>> the blink of an eye on the very last commitfest day ... and artifacts
>>> like
>>> this aren't doing anything to increase my confidence in it.

It was just missed comment after the fix between patch revisions.

> I really wish I could have done more to help with this, but I didn't
> do enough soon enough. Regrettably, I think that the patch just isn't
> ready. For example, the way that expression indexes just aren't
> handled is a cause for concern, as is the general way in which high
> keys are modified during index builds. Interactions with logical
> decoding are also a concern; there could be significant issues there,
> but that analysis just didn't happen. I had significant
> misunderstandings about the patch as recently as this week.
>
> This should be reverted.

The answer to the question about expressions is quite simple - they are not
supported by index-only scan, so having them in covering index now is just
wasting of disc space.

It's sad to hear that patch should be reverted. But, of course, I can't
argue with community's decision.

--
View this message in context: http://postgresql.nabble.com/pgsql-CREATE-INDEX-INCLUDING-column-tp5897653p5897721.html
Sent from the PostgreSQL - committers mailing list archive at Nabble.com.