| From: | Craig Ringer <craig(at)2ndquadrant(dot)com> |
|---|---|
| To: | Claudio Freire <klaussfreire(at)gmail(dot)com>, Merlin Moncure <mmoncure(at)gmail(dot)com> |
| Cc: | Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: possible optimization: push down aggregates |
| Date: | 2014-08-29 02:37:19 |
| Message-ID: | 53FFE75F.9040809@2ndquadrant.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 08/28/2014 03:46 AM, Claudio Freire wrote:
> You can't with mean and stddev, only with associative aggregates.
>
> That's min, max, sum, bit_and, bit_or, bool_and, bool_or, count.
You could with a new helper function to merge the temporary states for
each scan though.
In the case of mean, for example, it'd just mean adding the counts and sums.
However, I'm not sure how interesting that is without the ability to
execute the subplans in parallel.
--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2014-08-29 02:39:22 | Re: [BUGS] BUG #9652: inet types don't support min/max |
| Previous Message | Craig Ringer | 2014-08-29 02:17:17 | Re: Why data of timestamptz does not store value of timezone passed to it? |