From: | pilum(dot)70(at)uni-muenster(dot)de |
---|---|
To: | Peter Geoghegan <pg(at)heroku(dot)com> |
Cc: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: pg_stat_statements: calls under-estimation propagation |
Date: | 2013-10-10 11:50:23 |
Message-ID: | alpine.LNX.2.00.1310101313140.5867@ZIVPC313.uni-muenster.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Thx for your reply.
On Thu, 10 Oct 2013, Peter Geoghegan wrote:
> On Thu, Oct 10, 2013 at 3:11 AM, <pilum(dot)70(at)uni-muenster(dot)de> wrote:
>> But the drawback of this approach is impossibility to use
>> explain analyze without further substitutions.
>
> You can fairly easily disable the swapping of constants with '?'
> symbols, so that the query text stored would match the full originally
I thought I did ?! I introduced an additional user parameter
to disable the normalization in the patch shown in my last mail.
If there is already an easier way in the actual distribution,
i simply missed ist.
Where is this behaviour documented?
> executed query. Why would you want to, though? There could be many
> actual plans whose costs are aggregated as one query. Seeing one of
> them is not necessarily useful at all, and could be misleading.
>
Yeah, (thinking of for example parameter ranges) I mentioned that, I think,
but in the majority of cases beginners can easily conclude missing indices
executing explain analyze, because the queries, that are aggregated
and displayed under one query_id have very similar (or simply the same) query plans.
It's also only an option disabled by default: You can simply do nothing, if you don't like it :-)
VlG
Arne Scheffer
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2013-10-10 11:57:41 | Re: GIN improvements part 1: additional information |
Previous Message | Robert Haas | 2013-10-10 11:28:55 | Re: Auto-tuning work_mem and maintenance_work_mem |