planner's midjudge number of rows resulting, despite pretty obvious join

From: Grzegorz Jaśkiewicz <gryzman(at)gmail(dot)com>
To: "pgsql-performance(at)postgresql(dot)org" <pgsql-performance(at)postgresql(dot)org>
Subject: planner's midjudge number of rows resulting, despite pretty obvious join
Date: 2009-02-24 14:33:55
Message-ID: 2f4958ff0902240633y46e9ed96s3bc73434f75d269f@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

explain select ss, ARRAY(select id from foo where ss>0 and id between
7 and 156 order by random() limit 3) as v from
generate_series(1,1000000) ss;
QUERY PLAN
------------------------------------------------------------------------------------
Function Scan on generate_series ss (cost=0.00..9381.22 rows=1000 width=4)
SubPlan
-> Limit (cost=9.36..9.37 rows=3 width=8)
-> Sort (cost=9.36..9.74 rows=150 width=8)
Sort Key: (random())
-> Result (cost=0.00..7.42 rows=150 width=8)
One-Time Filter: ($0 > 0)
-> Seq Scan on foo (cost=0.00..7.05 rows=150 width=8)
Filter: ((id >= 7) AND (id <= 156))
(9 rows)

:(

no matter if I change last generate_series's range, it will always
estimate 1000 rows...

--
GJ

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Pavel Stehule 2009-02-24 14:43:01 Re: planner's midjudge number of rows resulting, despite pretty obvious join
Previous Message Claus Guttesen 2009-02-24 07:28:38 Re: Abnormal performance difference between Postgres and MySQL