| From: | David Rowley <david(dot)rowley(at)2ndquadrant(dot)com> |
|---|---|
| To: | Christophe Pettus <xof(at)thebuild(dot)com> |
| Cc: | "pgsql-hackers(at)postgresql(dot)org Hackers" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: HashAggregate row estimate = 200 |
| Date: | 2016-06-04 02:26:57 |
| Message-ID: | CAKJS1f9Re58SGE11r=GmnZO8PUHU-kj5vVsii6Nvi22h6WZtww@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 4 June 2016 at 14:10, Christophe Pettus <xof(at)thebuild(dot)com> wrote:
> Something I've noticed frequently is that HashAggregate will, especially if the children are Append with one of the nodes a non-seqscan, estimate 200 rows rather than a calculated vlaue. Where is that value coming from? The statistics target, a hardwired constant, or something else?
Most likely from DEFAULT_NUM_DISTINCT in selfuncs.h. This is the
fallback, when nothing else can be used.
--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Amit Kapila | 2016-06-04 03:12:04 | Re: XTM & parallel search |
| Previous Message | Masahiko Sawada | 2016-06-04 02:25:52 | Re: Reviewing freeze map code |