Message-ID: <3aa13399-8dac-e13a-5ce9-d528ab93efb8@enterprisedb.com>
Date: Thu, 7 Oct 2021 00:05:39 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.1.0
Subject: Re: using extended statistics to improve join estimates
Content-Language: en-US
To: Zhihong Yu <zyu@yugabyte.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
References: <c8c0ff31-3a8a-7562-bbd3-78b2ec65f16c@enterprisedb.com>
 <2ed75657-e084-9539-c6de-597e5675014c@enterprisedb.com>
 <ff412e82-5427-3b72-cf32-4700156dc47c@enterprisedb.com>
 <CALNJ-vRTL-5qjYrNEDkS=xg_mbDT6y98XGgD8jrJzcPLRSMSuw@mail.gmail.com>
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
In-Reply-To: 
 <CALNJ-vRTL-5qjYrNEDkS=xg_mbDT6y98XGgD8jrJzcPLRSMSuw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/3aa13399-8dac-e13a-5ce9-d528ab93efb8%40enterprisedb.com>
Precedence: bulk

On 10/6/21 23:03, Zhihong Yu wrote:
> Hi,
> 
> +       conditions2 = statext_determine_join_restrictions(root, rel, mcv);
> +
> +       /* if the new statistics covers more conditions, use it */
> +       if (list_length(conditions2) > list_length(conditions1))
> +       {
> +           mcv = stat;
> 
> It seems conditions2 is calculated using mcv, I wonder why mcv is 
> replaced by stat (for conditions1 whose length is shorter) ?
> 

Yeah, that's wrong - it should be the other way around, i.e.

     if (list_length(conditions1) > list_length(conditions2))

There's no test with multiple candidate statistics yet, so this went 
unnoticed :-/


regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company