Re: TRUNCATE ONLY with foreign keys and triggers disabled

On Mon, 14 Apr 2025, Adrian Klaver wrote:

> On 4/14/25 08:20, Dimitrios Apostolou wrote:
>>  On Mon, 14 Apr 2025, Adrian Klaver wrote:
>>
>>>  On 4/14/25 08:07, Laurenz Albe wrote:
>>>>   On Mon, 2025-04-14 at 17:05 +0200, Dimitrios Apostolou wrote:
>>>>>   I meant the *referencing* table has just been
>>>>>   populated. I'm trying to delete the *referenced* table and I get the
>>>>>   error.
>>>>
>>>>   That would break the foreign key constraint, right?
>>>>   PostgreSQL cannot allow that.
>>>
>>>  I believe the OP is disabling all triggers including system ones if I
>>>  follow
>>>  correctly and possibly running a foul of;
>>>
>>>  https://www.postgresql.org/docs/current/sql-altertable.html
>>>
>>>  " Disabling or enabling internally generated constraint triggers requires
>>>  superuser privileges; it should be done with caution since of course the
>>>  integrity of the constraint cannot be guaranteed if the triggers are not
>>>  executed."
>>
>>  Exactly that. I learned this from pg_restore --disable-triggers, as a way
>>  to speed-up insertion.
>>
>>  Since triggers are disabled, I assumed that postgresql shouldn't care
>>  about referential integrity in TRUNCATE.
>
> I'm going to say the depends on order of execution. Without seeing the code
> changes you have done there is really no way to determine what the exact
> issue is.

Sent patch here:

https://www.postgresql.org/message-id/flat/4589087c-ec6f-4407-1f82-6cb2e681ac0a%40gmx.net

The reason I first sent here (pgsql-general) is that I wondered if the
error is just an implementation shortcoming without deeper reasons, thus
it would be achievable to try to fix it myself.


Dimitris