Home > mailing lists

how does the planer to estimate row when i use order by and group by - Mailing list pgsql-performance

From	楊新波
Subject	how does the planer to estimate row when i use order by and group by
Date	August 12, 2014 05:59:15
Msg-id	CAK=TrEEHgo9Lwk3O5MQqqjUTX=21-zTyNaeqKdhhXyhBsKqhLA@mail.gmail.com Whole thread Raw
Responses	Re: how does the planer to estimate row when i use order by and group by (Marti Raudsepp <marti@juffo.org>)
List	pgsql-performance

hi ,everybody

why does the planer estimate 200 rows when i use order by and group by .

evn:postgresql 8.4 and 9.3

table:

CREATE TABLE a

(

id serial NOT NULL,

name character varying(20),

modifytime timestamp without time zone,

CONSTRAINT a_pk PRIMARY KEY (id)

)

SQL:

explain analyze

select * from

( select id from a order by id ) d

group by id;

Query plan:

"Group (cost=0.15..66.42 rows=200 width=4) (actual time=0.008..0.008 rows=0 loops=1)"

" -> Index Only Scan using a_pk on a (cost=0.15..56.30 rows=810 width=4) (actual time=0.006..0.006 rows=0 loops=1)"

" Heap Fetches: 0"

"Total runtime: 0.046 ms"

Can anybody suggest something or explain this behavior?

From: Maxim Boguk
Date: 09 August 2014, 07:22:48
Subject: Re: Optimization idea for long IN() lists

From: Mark Kirkwood
Date: 12 August 2014, 07:56:27
Subject: Re: 60 core performance with 9.3