Re: Rename of triggers for partitioned tables

From: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
To: Arne Roland <A(dot)Roland(at)index(dot)de>
Cc: vignesh C <vignesh21(at)gmail(dot)com>, Zhihong Yu <zyu(at)yugabyte(dot)com>, Pg Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Rename of triggers for partitioned tables
Date: 2021-07-20 01:39:04
Message-ID: 202107200139.5zceapxcs6z7@alvherre.pgsql
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On 2021-Jul-20, Arne Roland wrote:

> Is your patch based on master? It doesn't apply at my end.

It does ... master being dd498998a3 here,

$ patch -p1 < /tmp/renametrig-8.patch
patching file src/backend/commands/trigger.c
patching file src/backend/parser/gram.y
patching file src/test/regress/expected/triggers.out
patching file src/test/regress/sql/triggers.sql

applies fine.

>> I don't think we need to give a NOTICE when the trigger name does not
>> match; it doesn't really matter that the trigger was named differently
>> before the command, does it?

> I'd expect the command
> ALTER TRIGGER name ON table_name RENAME TO new_name;
> to rename a trigger named "name". We are referring the trigger via it's name after all. If a child is named differently we break with that assumption. I think informing the user about that, is very valuable.

Well, it does rename a trigger named 'name' on the table 'table_name',
as well as all its descendant triggers. I guess I am surprised that
anybody would rename a descendant trigger in the first place. I'm not
wedded to the decision of removing the NOTICE, though ... are there any
other votes for that, anyone?


Álvaro Herrera 39°49'30"S 73°17'W —

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message 2021-07-20 01:47:37 RE: Parallel INSERT SELECT take 2
Previous Message Masahiko Sawada 2021-07-20 01:25:52 Re: Skipping logical replication transactions on subscriber side