From: | Robert Haas <robertmhaas(at)gmail(dot)com> |
---|---|
To: | Ashutosh Bapat <ashutosh(dot)bapat(at)enterprisedb(dot)com> |
Cc: | Rushabh Lathia <rushabh(dot)lathia(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Print correct startup cost for the group aggregate. |
Date: | 2017-03-09 01:30:18 |
Message-ID: | CA+TgmobJFgq26mD3fjh6mVnyA7zLEHS09U2V4gYunjv6Vr0aTg@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Sun, Mar 5, 2017 at 11:32 PM, Ashutosh Bapat
<ashutosh(dot)bapat(at)enterprisedb(dot)com> wrote:
>> I think there have been
>> previous discussions of switching over to the practice for which you
>> are advocating here, but my impression (without researching) is that
>> the current practice is more like what Rushabh did.
>
> I am not sure Rushabh's approach is correct. Here's the excerpt from my mail.
>
>>> The reason the reason why startup_cost = input_startup_cost and not
>>> input_total_cost for aggregation by sorting is we don't need the whole
>>> input before the Group/Agg plan can produce the first row.
>
> With Rushabh's approach the startup cost of aggregation by sorting
> would be way higher that it's right now. Secondly, it would match that
> of hash aggregation and thus forcing hash aggregation to be chosen in
> almost all the cases.
You're right. I'm wrong. I take it all back.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2017-03-09 01:33:09 | Re: Cost model for parallel CREATE INDEX |
Previous Message | Andrew Dunstan | 2017-03-09 01:29:49 | Re: [COMMITTERS] pgsql: Use asynchronous connect API in libpqwalreceiver |