On Mon, Nov 1, 2021 at 7:28 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Mon, Nov 1, 2021 at 2:38 PM Greg Nancarrow <gregn4422@gmail.com> wrote:
> >
> > On Mon, Nov 1, 2021 at 5:07 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> > >
> > > I haven't followed the discussion on pg_publication_objects view but
> > > what is the primary use case of this view? If it's to list all tables
> > > published in a publication (e.g, "select * from pg_publication_objects
> > > where pubname = 'pub1'), pg_publication_objects view lacks the
> > > information of FOR ALL TABLES publications. And probably we can use
> > > pg_publication_tables instead. On the other hand, if it's to list all
> > > tables published in FOR ALL TABLES IN SCHEMA publications (e.g.,
> > > "select * from pg_publication_object where objtype = 'schema'), the
> > > view doesn't show tables published in such publications.
> > >
>
> Both the problems mentioned can be fixed if we follow the change
> suggested by me in one of the emails above [1].
>
> >
> > I think that Amit originally suggested to have a view that provides
> > information about the objects in each publication (like table, tables
> > in schema, sequence ...).
> >
>
> Right and I think as you also mentioned in your previous email it can
> save the effort of users if they want to know all the objects
> published via a publication.
Thank you for the explanation. Given we already have
pg_publication_tables view, if pg_publication_objects view also shows
all tables published in FOR ALL TABLES publications or FOR ALL TABLES
IN SCHEMA publications, there is essentially not much difference
between pg_publication_tables and pg_publication_objects except for
objtype column. Right? If so it'd be better to have one row for each
FOR ALL TABLES publication and FOR ALL TABLES IN SCHEMA publication
with objtype = 'database' or 'schema' etc, instead of individual
tables.
> I am just not sure if it is worth adding
> such a view or we leave it to users to find that information via
> querying individual views or system tables for objects.
I've not looked at the patch for logical replication of sequences but
the view becomes more useful once we support the new type of
replication object? If so, we can consider this view again after the
patch gets committed.
Regards,
--
Masahiko Sawada
EDB: https://www.enterprisedb.com/