Re: Performance improvement for joins where outer side is unique

From: Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp>
To: dgrowleyml(at)gmail(dot)com
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Performance improvement for joins where outer side is unique
Date: 2015-03-18 08:30:32
Message-ID: 20150318.173032.124713381.horiguchi.kyotaro@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

Hello, The attached is non-search version of unique join.

It is not fully examined but looks to work as expected. Many
small changes make the patch larger but affected area is rather
small.

What do you think about this?

> Hello, I don't have enough time for now but made some
> considerations on this.
>
> It might be a way that marking unique join peer at bottom and
> propagate up, not searching from top of join list.
> Around create_join_clause might be a candidate for it.
> I'll investigate that later.

regards,

--
Kyotaro Horiguchi
NTT Open Source Software Center

Attachment Content-Type Size
unique_join_horiguti_v0.patch text/x-patch 24.7 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Abhijit Menon-Sen 2015-03-18 09:53:00 Re: MD5 authentication needs help -SCRAM
Previous Message Andres Freund 2015-03-18 08:05:54 Re: Using 128-bit integers for sum, avg and statistics aggregates