Horton

Re: row filtering for logical replication

Dilip Kumar <dilipbalaut@gmail.com>

View thread
From: Dilip Kumar <dilipbalaut@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: "houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>, Alvaro Herrera <alvherre@alvh.no-ip.org>, Peter Smith <smithpb2250@gmail.com>, "tanghy.fnst@fujitsu.com" <tanghy.fnst@fujitsu.com>, Euler Taveira <euler@eulerto.com>, Greg Nancarrow <gregn4422@gmail.com>, vignesh C <vignesh21@gmail.com>, Ajin Cherian <itsajin@gmail.com>, Rahila Syed <rahilasyed90@gmail.com>, Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Önder Kalacı <onderkalaci@gmail.com>, japin <japinli@hotmail.com>, Michael Paquier <michael@paquier.xyz>, David Steele <david@pgmasters.net>, Craig Ringer <craig@2ndquadrant.com>, Amit Langote <amitlangote09@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-01-21T10:27:02Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Release cache tuple when no longer needed

  2. Add some additional tests for row filters in logical replication.

  3. Fix one of the tests introduced in commit 52e4f0cd47.

  4. Allow specifying row filters for logical replication of tables.

  5. Move scanint8() to numutils.c

  6. Replace Test::More plans with done_testing

  7. Reduce relcache access in WAL sender streaming logical changes

  8. Small cleanups related to PUBLICATION framework code

  9. Add a view to show the stats of subscription workers.

  10. Allow publishing the tables of schema.

  11. Doc: improve documentation of CREATE/ALTER SUBSCRIPTION.

  12. Add PublicationTable and PublicationRelInfo structs

  13. Remove unused argument "txn" in maybe_send_schema().

  14. Add prepare API support for streaming transactions in logical replication.

  15. Unify PostgresNode's new() and get_new_node() methods

  16. Use l*_node() family of functions where appropriate

  17. Add support for prepared transactions to built-in logical replication.

  18. Restore the portal-level snapshot after procedure COMMIT/ROLLBACK.

  19. Rename a parse node to be more general

  20. Remove unused column atttypmod from initial tablesync query

  21. SEARCH and CYCLE clauses

On Thu, Jan 20, 2022 at 4:54 PM Amit Kapila <amit.kapila16@gmail.com> wrote:

> + /*
> + * Unchanged toasted replica identity columns are only detoasted in the
> + * old tuple, copy this over to the new tuple.
> + */
> + if (att->attlen == -1 &&
> + VARATT_IS_EXTERNAL_ONDISK(tmp_new_slot->tts_values[i]) &&
> + !VARATT_IS_EXTERNAL_ONDISK(old_slot->tts_values[i]))
> + {
> + if (tmp_new_slot == new_slot)
> + {
> + tmp_new_slot = MakeSingleTupleTableSlot(desc, &TTSOpsVirtual);
> + ExecClearTuple(tmp_new_slot);
> + ExecCopySlot(tmp_new_slot, new_slot);
> + }
> +
> + tmp_new_slot->tts_values[i] = old_slot->tts_values[i];
> + tmp_new_slot->tts_isnull[i] = old_slot->tts_isnull[i];
> + }
> + }
>
>
> What is the need to assign new_slot to tmp_new_slot at the beginning
> of this part of the code? Can't we do this when we found some
> attribute that needs to be copied from the old tuple?
>
> The other part which is not clear to me by looking at this code and
> comments is how do we ensure that we cover all cases where the new
> tuple doesn't have values?
>

IMHO, the only part we are trying to handle is when the toasted attribute
is not modified in the new tuple.  And if we notice the update WAL the new
tuple is written as it is in the WAL which is getting inserted into the
heap page.  That means if it is external it can only be in
VARATT_IS_EXTERNAL_ONDISK format.  So I don't think we need to worry about
any intermediate format which we use for the in-memory tuples.  Sometimes
in reorder buffer we do use the INDIRECT format as well which internally
can store ON DISK format but we don't need to worry about that case either
because that is only true when we have the complete toast tuple as part of
the WAL and we recreate the tuple in memory in reorder buffer, so even if
it can by ON DISK format inside INDIRECT format but we have complete tuple.

-- 
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com