On Thu, Jul 18, 2024 at 3:33 PM David Rowley <dgrowleyml@gmail.com> wrote:
>
> On Thu, 18 Jul 2024 at 21:24, Ashutosh Bapat
> <ashutosh.bapat.oss@gmail.com> wrote:
> > If those GUCs are enabled, the planner consumes large amount of memory
> > and also takes longer irrespective of whether partitionwise plan is
> > used or not. That's why the default is false. If majority of those
> > joins use nested loop memory, or use index scans instead sorting,
> > memory consumption won't be as large. Saying that it "can" result in
> > large increase in execution memory is not accurate. But I agree that
> > we need to mention the effect of work_mem on partitionwise
> > join/aggregation.
>
> hmm? please tell me what word other than "can" best describes
> something that is possible to happen but does not always happen under
> all circumstances.
May I suggest "may"? :) [1], [2], [3].
My point is, we need to highlight the role of work_mem. So modify both
the descriptions.
[1] https://www.thesaurus.com/e/grammar/can-vs-may/
[2] https://www.britannica.com/dictionary/eb/qa/modal-verbs-may-might-can-could-and-ought
[3] https://www.merriam-webster.com/grammar/when-to-use-can-and-may
--
Best Wishes,
Ashutosh Bapat