Message-ID: <aee23e3a-19dc-95da-34a3-d34e070e4294@enterprisedb.com>
Date: Mon, 22 Nov 2021 00:37:32 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.2.0
Subject: Re: BUG #17295: Different query plan with Index Only Scan and Bitmap
 Index Scan.
Content-Language: en-US
To: udv.mail@gmail.com, pgsql-bugs@lists.postgresql.org,
 PG Bug reporting form <noreply@postgresql.org>
References: <17295-82a33e30e35c0cd5@postgresql.org>
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
In-Reply-To: <17295-82a33e30e35c0cd5@postgresql.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/aee23e3a-19dc-95da-34a3-d34e070e4294%40enterprisedb.com>
Precedence: bulk

Hi,

I'm not really sure which part you condides a bug? We simply don't 
translate between IN() and OR-ed conditions, and there may be some 
costing differences. So we may generate different plans, and the switch 
between index scan and bitmap index scan may happen at different points.

On 11/20/21 17:52, PG Bug reporting form wrote:
> 
> EXPLAIN SELECT * FROM t WHERE a IN (142,147,153,199);
> -- Bitmap Heap Scan on t  (cost=3739.74..50987.74 rows=200000 width=8)
> |
> --   Recheck Cond: (a = ANY ('{142,147,153,199}'::integer[]))
> |
> --   ->  Bitmap Index Scan on t_i  (cost=0.00..3689.74 rows=200000
> width=0)|
> --         Index Cond: (a = ANY ('{142,147,153,199}'::integer[]))
> |

This is rather strange, though. Why would this have estimate 200000? The 
other query has estimate 4, so it can't be because of missing stats etc. 
Or why would it fix after several minutes?


regards

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