Re: Proposal: Conflict log history table for Logical Replication

On Thu, Jul 2, 2026 at 3:33 PM shveta malik <shveta(dot)malik(at)gmail(dot)com> wrote:
>
> On Thu, Jul 2, 2026 at 1:39 PM Dilip Kumar <dilipbalaut(at)gmail(dot)com> wrote:
> >
> >
> > I rebased the remaining patches on top of HEAD. So far, I have run
> > pgindent and completed the doc merge for 0001. The 0002 and 0003 are
> > just rebased.
> >
>

I am focusing on 001 alone.
I was verifying conflict insertion for all conflict_types in regular
worker flow (not parallel), I have just one comment there:

For insert_exists conflict, both replica_identity (JSON) and
replica_identity_full (bool) are NULL. Is it conscious decision to
keep 'replica_identity_full' as NULL or shall we make this field as
'bool NOT NULL DEFAULT false'. As per doc at [1], NULL value for
boolean represents 'unknown'

If it is intentionaly kept NULL, do you think we should update below
doc to explain that it can be NULL for the cases where
replica_identity is not applicable.

replica_identity_full boolean I
indicates whether replica_identity represents a full tuple (true) or
key values of a replica identity index (false).

[1]: https://www.postgresql.org/docs/19/datatype-boolean.html

thanks
Shveta