Message-ID: <b1bd13be-e41c-461a-a969-1278183b0812@gmx.de>
Date: Wed, 10 Apr 2024 20:14:00 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Detoasting optionally to make Explain-Analyze less misleading
To: Tom Lane <tgl@sss.pgh.pa.us>,
 Matthias van de Meent <boekewurm+postgres@gmail.com>
Cc: Tomas Vondra <tomas.vondra@enterprisedb.com>, pgsql-hackers@postgresql.org
References: <ca0adb0e-fa4e-c37e-1cd7-91170b18cae1@gmx.de>
 <ce7edfd9-dcc5-2203-5ce5-051ca229b75a@gmx.de>
 <03f952d6-b773-be93-e5d5-8f9ea9c2d461@gmx.de>
 <7b7d46b1-d6f7-eb10-55c9-aa45fce57781@enterprisedb.com>
 <9126b7da-204b-4701-8e2b-37fb824a0cf4@gmx.de>
 <78507aa1-f850-50cd-458a-3986d973d092@enterprisedb.com>
 <CAEze2WjrZST8YEhV_oYXzC8y=P5aRGCMe_QpwhdK1PmNpjCeWQ@mail.gmail.com>
 <886b209f-8736-9cf2-ab9c-89870dea7bd8@enterprisedb.com>
 <CAEze2Wg8UMUuT9sHpWGCeBzm-nHTbQcnPsSob4ynAPXjiD-G=Q@mail.gmail.com>
 <14746b40-16a8-b53e-18a6-f2872e696e34@enterprisedb.com>
 <CAEze2WiLjoAYTcaaSR=pMK1SaZHMDh4Z7yeDeDrk3B7zNVA4gA@mail.gmail.com>
 <ea885631-21f1-425a-97ed-c4bfb8cf9c63@gmx.de>
 <CAEze2Wj=w+qDtPqM7qt-ZRXbG7vmvONa-ocfQZ1inGuE58zrRg@mail.gmail.com>
 <1642956.1712072848@sss.pgh.pa.us>
 <CAEze2WhopAFRS4xdNtma6XtYCRqydPWAg83jx8HZTowpeXzOyg@mail.gmail.com>
 <384054.1712181051@sss.pgh.pa.us> <CAEz!
 e2WhdmjbzVxgP8aW=0Dqynweh72QCdWRyxS24+1nmVopmrg@mail.gmail.com>
 <4017527.1712757455@sss.pgh.pa.us>
Content-Language: en-US
From: stepan rutz <stepan.rutz@gmx.de>
In-Reply-To: <4017527.1712757455@sss.pgh.pa.us>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: quoted-printable
UI-OutboundReport: notjunk:1;M01:P0:cubPFxKsc9A=;cl/ykh4/6Tf7WvsENVvn4kFo1fE
 W6pvqgTot7WT2XF+f0V/DWqG1y1DNn4J7c/dHAwIXnQs9BNa3mlIAz0WX5XLKgd3QerBREpqe
 aKNj2lD5NZetD/nVR4etvSH09K2ifZwT3WYdwlv+kMJxhG3wfYfakUfktvGutDV0xFKc4/BLr
 PIWf4X3PoV3LEUbCqV5oVH1ux8FfQyAnfmhLlENPBipndXa8t8X/kCKEQl+k9aY8LeAyTEV25
 P74GnsOKiF1dJvbXhNn0brWihwoAvw37slD/RLR7wFEGg9zd2nPh6cja9L8+91Pv8cDFGy9So
 2P1iLNRMyvdFw6xKBDbeAhkzoV1yQLJejTwKkxp0y5JhBJQwfy8q3CK52vO5OPpGR5G1r1uZ1
 MuoeYHkSI2B9ECtcT+k9gHdoR3+1UmKkM5r5TXe3OKD1hhDh6gYEePUsFcWbsEiYN/GJRstVW
 VIPoFVlmQezBmcJrVBIcfjB7xdz/aYB/QOovbq70AueroUdHidQy06pTzTEJibnHsze1ZKEBn
 jDIAmKlXv+e+EVzYVBRmZWG/om63MY7SDjNqawabItPYyFjitRV5ovcTRCde3tTIQPpwMMiH5
 NT//yqmPcoEgBvc86h/IRMZ/1A3CGz9UcpzjRxDCLwgfCQy+aIWq/o9acWACnywh1oObwwVB9
 jicS+BRYRztCHnJR3IOYte9bJdfnrtA1lWtnO5mbOoKpgMqq0UXjZsPxjT2o4N8bKV8FeOQL4
 F0FzEr5reRMMrBJseEbOmG+fPq9/KY9Sta+5eikPHpyjwOebOFC521qJbTUaVgTIzAZTYoBvX
 IvOP0GfXCKUDs1skl9YFLGIZChaMVmteXIP0R3JyTt3wM=
Archived-At: 
 <https://www.postgresql.org/message-id/b1bd13be-e41c-461a-a969-1278183b0812%40gmx.de>
Precedence: bulk

First of all thanks for bringing this Feature to PostgreSQL. From a
regular-user perspective (not everyone is a Pro) it is very misleading
that ANALYZE doesn't do what it suggests it does. To run the query into
some kind of /dev/null type of destination is feasible and that is what
people end up doing after they have fallen into the "de-toasting" trap.

Having SERIALIZE is a great improvement for certain. When I said that
SERIALIZE should be the default, then this came mostly out of very
surprising subjective experiences in the past. Turning it on certainly
alters some existing benchmarks and timings. That is destructive in a
way and would destroy some existing work and measures. I lack the
overall understanding of the consequences, so please don't follow this
(emotional) advice.

So thanks again! and this will really help a lot of people. The people
actually bothering with EXPLAIN options are likely to explore the
documentation and now have a hint about this pitfall. The EXPLAIN part
of PostgreSQL "feels" a lot better now.

I appreciate all of your work on this issue, which came up without being
on some kind of plan and of course for the overall work on PostgreSQL.

/Stepan

On 4/10/24 15:57, Tom Lane wrote:

> Matthias van de Meent <boekewurm+postgres@gmail.com> writes:
>> Upthread at [0], Stepan mentioned that we should default to SERIALIZE
>> when ANALYZE is enabled. I suspect a patch in that direction would
>> primarily contain updates in the test plan outputs, but I've not yet
>> worked on that.
>> Does anyone else have a strong opinion for or against adding SERIALIZE
>> to the default set of explain features enabled with ANALYZE?
> I am 100% dead set against that, because it would silently render
> EXPLAIN outputs from different versions quite non-comparable.
>
> 			regards, tom lane