Re: BUG #18103: bugs of concurrent merge into when use different join plan - Mailing list pgsql-bugs

From Dean Rasheed
Subject Re: BUG #18103: bugs of concurrent merge into when use different join plan
Date
Msg-id CAEZATCVGWykA48zdMNziD=u0cipioYKTKy_+7S_UBxM828kYAg@mail.gmail.com
Whole thread Raw
In response to Re: BUG #18103: bugs of concurrent merge into when use different join plan  (Richard Guo <guofenglinux@gmail.com>)
List pgsql-bugs
On Tue, 26 Sept 2023 at 10:46, Richard Guo <guofenglinux@gmail.com> wrote:
>
> Agreed.  We need rowmarks for MERGE too.  I found that the following
> comment from plannodes.h is very useful in understanding this.
>
> * When doing UPDATE, DELETE, or SELECT FOR UPDATE/SHARE, we have to uniquely
> * identify all the source rows, not only those from the target relations, so
> * that we can perform EvalPlanQual rechecking at need.
>

Ah, good. Thanks. That confirms my understanding.

> To nitpick, there are some comments that may need to be updated to
> include MERGE, such as the comments for ExecRowMark, RowMarkType and
> PlanRowMark.
>

Agreed. I'll see about doing that.

Regards,
Dean



pgsql-bugs by date:

Previous
From: Kyle MacMillan
Date:
Subject: md5 password valid and invalid after upgrading
Next
From: Laurenz Albe
Date:
Subject: Re: md5 password valid and invalid after upgrading