merge semi join cost calculation error

From: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
To: PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: merge semi join cost calculation error
Date: 2018-10-08 14:26:37
Message-ID: CAFj8pRDnixz78Pw-S_yACaA9hSXGGUDrmPX0G4U4Vaz8itF39g@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi

I try to understand to a issue
https://stackoverflow.com/questions/52685384/subquery-performance-on-simple-case

The user sent a plan:

QUERY PLAN
Merge Semi Join (cost=82.97..580.24 rows=580 width=8) (actual
time=0.503..9557.396 rows=721 loops=1)
Merge Cond: (tips.users_id = follows.users_id_to)
-> Index Scan using tips_idx_users_id01 on tips (cost=0.43..8378397.19
rows=2491358 width=16) (actual time=0.009..9231.585 rows=2353914 loops=1)
-> Sort (cost=1.77..1.82 rows=22 width=8) (actual time=0.052..0.089
rows=28 loops=1)
Sort Key: follows.users_id_to
Sort Method: quicksort Memory: 26kB
-> Seq Scan on follows (cost=0.00..1.27 rows=22 width=8) (actual
time=0.013..0.020 rows=28 loops=1)
Filter: (users_id_from = 1)

He has PostgreSQL 10.5. I cannot to understand to too low total cost of Merge
Semi Join because subnode has very high cost 8378397.

I cannot to emulate this case on my comp - so it looks like maybe some
build error. What do you think about?

Regards

Pavel

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2018-10-08 14:27:19 Re: chained transactions
Previous Message Andrew Dunstan 2018-10-08 13:55:51 Re: exclude tmp_check and tmp_install from pgindent