Re: ALTER COLUMN ... SET EXPRESSION to alter stored generated column's expression - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: ALTER COLUMN ... SET EXPRESSION to alter stored generated column's expression
Date
Msg-id fbc544ee-5c16-dff2-0f74-bcffb0a0aa1d@eisentraut.org
Whole thread Raw
In response to Re: ALTER COLUMN ... SET EXPRESSION to alter stored generated column's expression  (Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>)
Responses Re: ALTER COLUMN ... SET EXPRESSION to alter stored generated column's expression
List pgsql-hackers
On 06.10.23 14:57, Ashutosh Bapat wrote:
> On Fri, Oct 6, 2023 at 6:06 PM Peter Eisentraut <peter@eisentraut.org> wrote:
>>
>> On 28.08.23 11:54, Amul Sul wrote:
>>> Thanks for the review comments, I have fixed those in the attached
>>> version. In
>>> addition to that, extended syntax to have the STORE keyword as suggested by
>>> Vik.
>>
>> An additional comment: When you change the generation expression, you
>> need to run ON UPDATE triggers on all rows, if there are any triggers
>> defined.  That is because someone could have triggers defined on the
>> column to either check for valid values or propagate values somewhere
>> else, and if the expression changes, that is kind of like an UPDATE.
>>
>> Similarly, I think we should consider how logical decoding should handle
>> this operation.  I'd imagine it should generate UPDATE events on all
>> rows.  A test case in test_decoding would be useful.
> 
> Should we treat it the same fashion as ALTER COLUMN ... TYPE which
> rewrites the column values? Of course that rewrites the whole table,
> but logically they are comparable.

I don't know.  What are the semantics of that command with respect to 
triggers and logical decoding?




pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Add const to values and nulls arguments
Next
From: Peter Eisentraut
Date:
Subject: Re: Allow tests to pass in OpenSSL FIPS mode