Thread: Can somebody explain what is the meaning for HashAggregate?

Can somebody explain what is the meaning for HashAggregate?

From
lin
Date:
Can somebody explain what does the postgres done for the explain of sql shows HashAggregate( what is the meaning for  HashAggregate  )?
for example: 

postgres=# explain verbose select oid,relname  from pg_class group by oid,relname;
                                 QUERY PLAN                                  
-----------------------------------------------------------------------------
 HashAggregate  (cost=12.42..15.38 rows=295 width=68)
   Output: oid, relname
   Group Key: pg_class.oid, pg_class.relname
   ->  Seq Scan on pg_catalog.pg_class  (cost=0.00..10.95 rows=295 width=68)
         Output: oid, relname
(5 rows)

::  first, seq scan pg_class and return (oid,relname);
   second,  make group (oid,relname),  is it fisrt sort by oid then in the oid group sort by relname?
Can somebody explain what does the database done for hashAggregate?

 Thanks,
  wanglin


 

Re: Can somebody explain what is the meaning for HashAggregate?

From
Bill Moran
Date:
On Sun, 20 Sep 2015 21:56:39 +0800 (CST)
lin <jluwln@163.com> wrote:

> Can somebody explain what does the postgres done for the explain of sql shows HashAggregate( what is the meaning for
HashAggregate )? 
> for example:
>
>
> postgres=# explain verbose select oid,relname  from pg_class group by oid,relname;
>                                  QUERY PLAN
> -----------------------------------------------------------------------------
>  HashAggregate  (cost=12.42..15.38 rows=295 width=68)
>    Output: oid, relname
>    Group Key: pg_class.oid, pg_class.relname
>    ->  Seq Scan on pg_catalog.pg_class  (cost=0.00..10.95 rows=295 width=68)
>          Output: oid, relname
> (5 rows)
>
>
> ::  first, seq scan pg_class and return (oid,relname);
>    second,  make group (oid,relname),  is it fisrt sort by oid then in the oid group sort by relname?
> Can somebody explain what does the database done for hashAggregate?

It combines the values for oid and relname for each returned row, generates a hashkey
for them, then uses that hashkey to aggregate (compute the GROUP BY, essentially, in
this case)

--
Bill Moran