Re: BUG #19099: Conditional DELETE from partitioned table with non-updatable partition raises internal error

amit <amitlangote09@gmail.com>

From: Amit Langote <amitlangote09@gmail.com>

To: Tender Wang <tndrwang@gmail.com>

Cc: David Rowley <dgrowleyml@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Kirill Reshke <reshkekirill@gmail.com>, jian he <jian.universality@gmail.com>, exclusion@gmail.com, pgsql-bugs@lists.postgresql.org

Date: 2025-11-07T06:03:50Z

Lists: pgsql-bugs

Hi,

On Fri, Nov 7, 2025 at 10:01 AM Tender Wang <tndrwang@gmail.com> wrote:
> Amit Langote <amitlangote09@gmail.com> 于2025年11月6日周四 18:00写道：
>> So for back branches, I’d propose sticking with the smaller
>> executor-side fix and perhaps revisiting the planner behavior
>> separately if we ever want to refine handling of pruned partitions or
>> dummy roots. I understand, as was reported upthread, that the EXPLAIN
>> VERBOSE output isn’t very consistent with that patch even though the
>> internal error goes away.  Making sense of the output differences
>> requires knowing that the targetlist population behavior differs
>> depending on whether enable_partition_pruning is on or off as I
>> described above.
>
> The executor-side fix works for me

Thanks for checking.

> and the test case should be added to your patch.
> Should we add some comments to explain the output difference in EXPLAIN VERBOSE
> if enable_partition_pruning is set to a different value?

I added some in the v2 patch I just posted.

-- 
Thanks, Amit Langote