Horton

Re: Dumping/restoring fails on inherited generated column

Anastasia Lubennikova <a.lubennikova@postgrespro.ru>

View thread View on postgr.es
From: Anastasia Lubennikova <a.lubennikova@postgrespro.ru>
To: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>, Masahiko Sawada <sawada.mshk@gmail.com>
Cc: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2020-11-30T15:37:22Z
Lists: pgsql-hackers 
On 09.11.2020 13:43, Peter Eisentraut wrote:
> On 2020-11-06 04:55, Masahiko Sawada wrote:
>>> Both of these result in the same change to the dump output.  Both of
>>> them have essentially the same idea.  The first one adds the
>>> conditionals during the information gathering phase of pg_dump, the
>>> second one adds the conditionals during the output phase.
>>>
>>> Any further thoughts?
>> I think the first one is better than the second (mine) because it can
>> save the number of intermediate objects.
>
> I was hoping to wrap this issue up this week, but I found more 
> problems with how these proposed changes interact with 
> --binary-upgrade mode.  I think I need to formalize my findings into 
> pg_dump test cases as a next step.  Then we can figure out what 
> combination of tweaks will make them all work.
>
I am moving this patch to the next CF, but it looks like the discussion 
is a bit stuck.


Peter, can you please share your concerns about the interaction of the 
patch with --binary-upgrade mode? If you don't have time to write tests, 
you can just describe problems.

-- 
Anastasia Lubennikova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

Commits

  1. Fix ALTER TABLE / INHERIT with generated columns

  2. pg_dump: Fix dumping of inherited generated columns

  3. Disallow ALTER TABLE ONLY / DROP EXPRESSION

  4. Fix several DDL issues of generated columns versus inheritance