From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Benoit Lobréau <benoit(dot)lobreau(at)gmail(dot)com> |
Cc: | pgsql-general <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Explain analyse and toasted data. |
Date: | 2017-08-24 12:18:17 |
Message-ID: | 31374.1503577097@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
=?UTF-8?Q?Benoit_Lobr=C3=A9au?= <benoit(dot)lobreau(at)gmail(dot)com> writes:
> In this video (PostgreSQL at 10 TB and beyond: https://youtu.be/8mKpfutwD0U
> at 6mn05), the speaker says that on very big databases explain analyse can
> give unrealistic results because it doesn't read toasted data.
> I suppose it's because of this: "The big values of TOASTed attributes will
> only be pulled out (if selected at all) at the time the result set is sent
> to the client." But I dont know if explain analyse does something to
> simulate this (if it's possible).
EXPLAIN ANALYZE does not do that. It likewise does not do anything to try
to model the network transmission costs, which are also likely to be
significant if the data is bulky --- but there's no way to do that without
actually sending the result data to the client, AFAICS.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2017-08-24 12:51:36 | Re: Retrieving query results |
Previous Message | Tom Lane | 2017-08-24 11:56:52 | Re: DROP [TEMP] TABLE syntax, as reason why not? |