Horton

Re: pg_stat_statements and "IN" conditions

Dmitry Dolgov <9erthalion6@gmail.com>

View thread
From: Dmitry Dolgov <9erthalion6@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Zhihong Yu <zyu@yugabyte.com>, David Steele <david@pgmasters.net>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Greg Stark <stark@mit.edu>, Pavel Trukhanov <pavel.trukhanov@gmail.com>
Date: 2022-03-10T16:38:37Z
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. Introduce squashing of constant lists in query jumbling

  2. Make documentation builds reproducible

  3. Include values of A_Const nodes in query jumbling

  4. Teach planner about more monotonic window functions

  5. Split up guc.c for better build speed and ease of maintenance.

> On Wed, Jan 05, 2022 at 10:11:11PM +0100, Dmitry Dolgov wrote:
> > On Tue, Jan 04, 2022 at 06:02:43PM -0500, Tom Lane wrote:
> > We can debate whether the rules proposed here are good for
> > pg_stat_statements or not, but it seems inevitable that they will be a
> > disaster for some other consumers of the query hash.
>
> Hm, which consumers do you mean here, potential extension? Isn't the
> ability to use an external module to compute queryid make this situation
> possible anyway?
>
> > do you really think that a query with two int
> > parameters is equivalent to one with five float parameters for all
> > query-identifying purposes?
>
> Nope, and it will be hard to figure this out no matter which approach
> we're talking about, because it mostly depends on the context and type
> of queries I guess. Instead, such functionality should allow some
> reasonable configuration. To be clear, the use case I have in mind here
> is not four or five, but rather a couple of hundreds constants where
> chances that the whole construction was generated automatically by ORM
> is higher than normal.
>
> > I can see the merits of allowing different numbers of IN elements
> > to be considered equivalent for pg_stat_statements, but this patch
> > seems to go far beyond that basic idea, and I fear the side-effects
> > will be very bad.
>
> Not sure why it goes far beyond, but then there were two approaches
> under consideration, as I've stated in the first message. I already
> don't remember all the details, but another one was evolving around
> doing similar things in a more limited fashion in transformAExprIn. The
> problem would be then to carry the information, necessary to represent
> the act of "merging" some number of queryids together. Any thoughts
> here?
>
> The idea of keeping the original queryid untouched and add another type
> of id instead sounds interesting, but it will add too much overhead for
> a quite small use case I guess.

```
Thu, 10 Mar 2022
New status: Waiting on Author
```

This seems incorrect, as the only feedback I've got was "this is a bad
idea", and no reaction on follow-up questions.