Query choosing Bad Index Path - Mailing list pgsql-performance

From Valli Annamalai
Subject Query choosing Bad Index Path
Date
Msg-id CADkhgiJ+gT_FDKZWgP8oZsy6iRbYMYkmRjsPhqhcT1A2KBgcHA@mail.gmail.com
Whole thread Raw
Responses Re: Query choosing Bad Index Path
List pgsql-performance

Postgres version: 11.4
Problem:
    Query choosing Bad Index Path. Details are provided below:


Table :






Doubt
   1. Why is this Query choosing Index Scan Backward using table1_pkey Index though it's cost is high. It can rather choose
            BITMAP OR
                  (Index on RECORDID) i.e; table1_idx6
                  (Index on RELATEDID) i.e; table1_idx7

      Below is the selectivity details from pg_stats table
        - Recordid has 51969 distinct values. And selectivity (most_common_freqs) for recordid = 15842006928391817 is 0.00376667
        - Relatedid has 82128 distinct values. And selectivity (most_common_freqs) for recordid = 15842006928391817 is 0.0050666

Since, selectivity is less, this should logically choose this Index, which would have improve my query performance here.

I cross-checked the same by removing PrimaryKey to this table and query now chooses these indexes and response is in 100ms. Please refer the plan below (after removing primary key):
















Attachment

pgsql-performance by date:

Previous
From: Vijaykumar Jain
Date:
Subject: Re: Terribly slow query with very good plan?
Next
From: Pavel Stehule
Date:
Subject: Re: Query choosing Bad Index Path