Re: row filtering for logical replication

On Wed, Sep 22, 2021 at 6:42 AM Ajin Cherian <itsajin(at)gmail(dot)com> wrote:
>
> On Tue, Sep 21, 2021 at 9:42 PM Dilip Kumar <dilipbalaut(at)gmail(dot)com> wrote:
> >
> > On Tue, Sep 21, 2021 at 4:29 PM Dilip Kumar <dilipbalaut(at)gmail(dot)com> wrote:
> >
> > I have one more question, while looking into the
> > ExtractReplicaIdentity() function, it seems that if any of the "rep
> > ident key" fields is changed then we will write all the key fields in
> > the WAL as part of the old tuple, not just the changed fields. That
> > means either the old tuple will be NULL or it will be having all the
> > key attributes. So if we are supporting filter only on the "rep ident
> > key fields" then is there any need to copy the fields from the new
> > tuple to the old tuple?
> >
>
> Yes, I just figured this out while testing. So we don't need to copy fields
> from the new tuple to the old tuple.
>
> But there is still the case of your fix for the unchanged toasted RI
> key fields in the new tuple
> which needs to be copied from the old tuple to the new tuple. This
> particular case
> seems to violate both rules that an old tuple will be present only
> when there are changed
> RI key fields and that if there is an old tuple it will contain all RI
> key fields.
>

Why do you think that the second assumption (if there is an old tuple
it will contain all RI key fields.) is broken? It seems to me even
when we are planning to include unchanged toast as part of old_key, it
will contain all the key columns, isn't that true?

> I think we
> still need to deform both old tuple and new tuple, just to handle this case.
>

Yeah, but we will anyway talking about saving that cost for later if
we decide to send that tuple. I think we can further try to optimize
it by first checking whether the new tuple has any toasted value, if
so then only we need this extra pass of deforming.

> There is currently logic in ReorderBufferToastReplace() which already
> deforms the new tuple
> to detoast changed toasted fields in the new tuple. I think if we can
> enhance this logic for our
> purpose, then we can avoid an extra deform of the new tuple.
> But I think you had earlier indicated that having untoasted unchanged
> values in the new tuple
> can be bothersome.
>

I think it will be too costly on the subscriber side during apply
because it will update all the unchanged toasted values which will
lead to extra writes both for WAL and data.

--
With Regards,
Amit Kapila.