Horton

Re: Early WIP/PoC for inlining CTEs

Andres Freund <andres@anarazel.de>

View thread
From: Andres Freund <andres@anarazel.de>
To: Thomas Munro <thomas.munro@enterprisedb.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Andreas Karlsson <andreas@proxel.se>, Andrew Gierth <andrew@tao11.riddles.org.uk>, Tom Lane <tgl@sss.pgh.pa.us>, David Fetter <david@fetter.org>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2019-01-11T19:04:03Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Prevent inlining of multiply-referenced CTEs with outer recursive refs.

  2. Allow user control of CTE materialization, and change the default behavior.

  3. Split QTW_EXAMINE_RTES flag into QTW_EXAMINE_RTES_BEFORE/_AFTER.

  4. document when PREPARE uses generic plans

Hi,

On 2019-01-12 07:58:39 +1300, Thomas Munro wrote:
> On Sat, Jan 12, 2019 at 7:19 AM Andres Freund <andres@anarazel.de> wrote:
> > On 2019-01-11 11:12:25 -0500, Robert Haas wrote:
> > > I actually think that we should go "all in" here and allow the user to
> > > specify either that they want materialization or that they don't want
> > > materialization.  If they specify neither, then we make some decision
> > > which we may change in a future release.  If they do specify
> > > something, we do that.
> >
> > +many
> 
> I think the syntax as proposed is almost OK if we only want to
> grandfather in our historically hintful CTEs but discourage the
> development of any future kinds of hints.  Even then I don't love the
> way it formalises a semi-procedural step at the same language level as
> a glorious declarative relational query.
> 
> Maybe we could consider a more extensible syntax that is attached to
> the contained SELECT rather than the containing WITH.  Then CTEs would
> be less special; there'd be a place to put hints controlling top-level
> queries, subselects, views etc too (perhaps eventually join hints,
> parallelism hints etc, but "materialize this" would be just another
> one of those things).  That'd be all-in.

I think you have some purity arguments here, but the likelihood of us
developing a full-blown solution is not that high, and the lack of
inlinable CTEs is *really* hurting us. As long as the design doesn't
block a full solution, if we go there, I think it's a very acceptable
blemish in comparison to the benefits we'd get.

Greetings,

Andres Freund