Re: [PATCH] Negative Transition Aggregate Functions (WIP)

From: Vladimir Sitnikov <sitnikov(dot)vladimir(at)gmail(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Florian Pflug <fgp(at)phlo(dot)org>, Greg Stark <stark(at)mit(dot)edu>, David Rowley <dgrowleyml(at)gmail(dot)com>, Josh Berkus <josh(at)agliodbs(dot)com>, Kevin Grittner <kgrittn(at)ymail(dot)com>, Pgsql Hackers <pgsql-hackers(at)postgresql(dot)org>, Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com>
Subject: Re: [PATCH] Negative Transition Aggregate Functions (WIP)
Date: 2014-03-05 18:40:59
Message-ID: CAB=Je-GmXuinuFrfFBykf=1XRMjygKbUq9_-evPJcmej6RFo6Q@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Tom,

I did not follow the thread very close, so I need to look what the
ambiguity is there, however I would love to see window rescans in explain
analyze.

I have great experience in tuning Oracle queries.
There are features in PostgreSQL's explain analyze that I miss badly in
Oracle: 'rows removed by filter' is my favourite one. Improving explain
analyze is great for performance analysis.

I would say target audience for 'explain analyze' is not all the users, but
someone closer to 'performance engineers'. Those beings are used to
triple-check results and build/validate hypothesis since all the counters
tend to lie, so 'a bit misleading counter' is not a showstopper.

I did not think of
'non-being-able-to-use-negative-transition-since-floats-do-not-commute'
before.
Thanks to this discussion I see what kind of dragons live here.

I would vote (if I had any vote at all) for the inclusion of performance
statistics to explain analyze (e.g. number of rescans and number of rows
fed to aggregate) provided performance impact is tolerable in both regular
and explain analyze mode.

I wonder how Oracle handles negative transition (does it?), however that is
a different discussion.

Regards,
Vladimir Sitnikov

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2014-03-05 18:49:23 Re: Changeset Extraction v7.9.1
Previous Message Josh Berkus 2014-03-05 18:05:37 Re: jsonb and nested hstore