Horton

Re: Default setting for enable_hashagg_disk

David Rowley <dgrowleyml@gmail.com>

View thread View on postgr.es
From: David Rowley <dgrowleyml@gmail.com>
To: "David G. Johnston" <david.g.johnston@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Stephen Frost <sfrost@snowman.net>, Peter Geoghegan <pg@bowt.ie>, Jeff Davis <pgsql@j-davis.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, Andres Freund <andres@anarazel.de>, Bruce Momjian <bruce@momjian.us>, Robert Haas <robertmhaas@gmail.com>, Justin Pryzby <pryzby@telsasoft.com>, Melanie Plageman <melanieplageman@gmail.com>, Tomas Vondra <tomas.vondra@2ndquadrant.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2020-07-11T01:43:28Z
Lists: pgsql-hackers, pgsql-docs 
On Sat, 11 Jul 2020 at 13:36, David G. Johnston
<david.g.johnston@gmail.com> wrote:
> If we add a setting that defaults to work_mem then the benefit is severely reduced.  You still have to modify individual queries, but the change can simply be more targeted than changing work_mem alone.

I think the idea is that this is an escape hatch to allow users to get
something closer to what PG12 did, but only if they really need it.  I
can't quite understand why we need to leave the escape hatch open and
push them halfway through it.  I find escape hatches are best left
closed until you really have no choice but to use them.

David

Commits

  1. Add hash_mem_multiplier GUC.

  2. HashAgg: use better cardinality estimate for recursive spilling.

  3. Remove hashagg_avoid_disk_plan GUC.

  4. Doc fixup for hashagg_avoid_disk_plan GUC.

  5. Rework HashAgg GUCs.

  6. Disk-based Hash Aggregation.

  7. Implement partition-wise grouping/aggregation.

  8. Defer creation of partially-grouped relation until it's needed.