MIME-Version: 1.0
References: 
 <CAOKkKFvYHSEsFazkrf9bRH14p-H27XMaqbZfRYjS6EHBruvZMQ@mail.gmail.com>
 <CAExHW5tPhCw5HpHG6F7uAXx3re_gRXqm_ryJFjFQ2+WrVY+hPQ@mail.gmail.com>
 <CAOKkKFsLbbpVMCG2_bV8u7M2sXLTs0y_yfMpywKekGi+qKtWKw@mail.gmail.com>
 <CAOKkKFvgjAwtUFKh3baZ7BcQ5u+wS_DO+2n7dh0un+19v_VOzQ@mail.gmail.com>
 <CAApHDvqq21KL9-eCfUyAUqxEyz40dut8+ehQHJD3Rh75iEMycw@mail.gmail.com>
 <CAOKkKFvBi5tNaPkP-VEUcAr=nek0nm9=sUnP+dmqCm-vmnL9DA@mail.gmail.com>
 <CAMbWs4-rqGC78jHduXO_feXCFzK=gzrcSHRTkATTKChG3YGq0A@mail.gmail.com>
 <TY4P301MB129931A356F6934C432BB0259EEC2@TY4P301MB1299.JPNP301.PROD.OUTLOOK.COM>
In-Reply-To: 
 <TY4P301MB129931A356F6934C432BB0259EEC2@TY4P301MB1299.JPNP301.PROD.OUTLOOK.COM>
From: Richard Guo <guofenglinux@gmail.com>
Date: Fri, 26 Jun 2026 15:26:20 +0900
Message-ID: 
 <CAMbWs48cWQ=QkwH21zhfGH2MsAFYJpskjak99N4VX2yB69k-Eg@mail.gmail.com>
Subject: Re: Add enable_groupagg GUC parameter to control GroupAggregate usage
To: Tatsuro Yamada <tatsuro.yamada@ntt.com>
Cc: Tatsuro Yamada <yamatattsu@gmail.com>,
 David Rowley <dgrowleyml@gmail.com>,
	Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>,
	"pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAMbWs48cWQ%3DQkwH21zhfGH2MsAFYJpskjak99N4VX2yB69k-Eg%40mail.gmail.com>
Precedence: bulk

On Thu, Jun 25, 2026 at 6:12=E2=80=AFPM Tatsuro Yamada <tatsuro.yamada@ntt.=
com> wrote:
> It sounds like there are several plan nodes that could be covered by this=
 GUC
> (such as SetOp, sort-based Unique for DISTINCT, semijoin uniqueification,
> and Group nodes). Do you think we should cover all of them before the pat=
ch
> would be considered complete enough for commit?

Yeah, I think so.  The v4 patch actually already did that.  You can
find these changes in create_setop_path(), create_unique_path(), and
cost_group().

> I'd be interested to hear your thoughts on how best to proceed, and wheth=
er
> dividing the work into smaller pieces would make sense.

It seems that the scope of the regression test changes isn't too
large, so I think keeping them in one patch is fine.

- Richard