Re: Overhead of union versus union all - Mailing list pgsql-general

From Adam Rich
Subject Re: Overhead of union versus union all
Date
Msg-id 4A564F0B.9090201@sbcglobal.net
Whole thread Raw
In response to Overhead of union versus union all  (Tim Keitt <tkeitt@keittlab.org>)
List pgsql-general
Tim Keitt wrote:
> I am combining query results that I know are disjoint. I'm wondering
> how much overhead there is in calling union versus union all. (Just
> curious really; I can't see a reason not to use union all.) (cc me
> please; not subscribed...)
>
> THK
>


I think you can test this one yourself pretty easily.  Just run the two
queries with "explain analyze".  Union All should run in about the sum
of the separate queries.  Plain Union will always be slower, because it
takes the same results from "union all" and runs them through an extra
sort/distinct or hash step.  In my tests, on a query with 600,000 rows,
the Plain Union took about 3x as long to complete.



pgsql-general by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: Overhead of union versus union all
Next
From: Bruce Momjian
Date:
Subject: Re: Overhead of union versus union all