Horton

RE: row filtering for logical replication

Zhijie Hou (Fujitsu) <houzj.fnst@fujitsu.com>

View thread
From: "houzj.fnst@fujitsu.com" <houzj.fnst@fujitsu.com>
To: Ajin Cherian <itsajin@gmail.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>, Euler Taveira <euler@eulerto.com>, Greg Nancarrow <gregn4422@gmail.com>, vignesh C <vignesh21@gmail.com>, Peter Smith <smithpb2250@gmail.com>, "tanghy.fnst@fujitsu.com" <tanghy.fnst@fujitsu.com>, Dilip Kumar <dilipbalaut@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: 2021-12-09T02:37:14Z
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 Wednesday, December 8, 2021 7:52 PM Ajin Cherian <itsajin@gmail.com>
> On Tue, Dec 7, 2021 at 5:36 PM Peter Smith <smithpb2250@gmail.com> wrote:
> >
> > We were mid-way putting together the next v45* when your latest
> > attachment was posted over the weekend. So we will proceed with our
> > original plan to post our v45* (tomorrow).
> >
> > After v45* is posted we will pause to find what are all the
> > differences between your unified patch and our v45* patch set. Our
> > intention is to integrate as many improvements as possible from your
> > changes into the v46* etc that will follow tomorrow’s v45*. On some
> > points, we will most likely need further discussion.
> 
> 
> Posting an update for review comments, using contributions majorly from
> Peter Smith.
> I've also included changes based on Euler's combined patch, specially changes
> to documentation and test cases.
> I have left out Hou-san's 0005, in this patch-set. Hou-san will provide a rebased
> update based on this.
> 
> This patch addresses the following review comments:

Hi,

Thanks for updating the patch.
I noticed a possible issue.

+				/* Check row filter. */
+				if (!pgoutput_row_filter(data, relation, oldtuple, NULL, relentry))
+					break;
+
+				maybe_send_schema(ctx, change, relation, relentry);
+
 				/* Switch relation if publishing via root. */
 				if (relentry->publish_as_relid != RelationGetRelid(relation))
 				{
...
					/* Convert tuple if needed. */
					if (relentry->map)
						tuple = execute_attr_map_tuple(tuple, relentry->map);

Currently, we execute the row filter before converting the tuple, I think it could
get wrong result if we are executing a parent table's row filter and the column
order of the parent table is different from the child table. For example:

----
create table parent(a int primary key, b int) partition by range (a);
create table child (b int, a int primary key);
alter table parent attach partition child default;
create publication pub for table parent where(a>10) with(PUBLISH_VIA_PARTITION_ROOT);

The column number of 'a' is '1' in filter expression while column 'a' is the
second one in the original tuple. I think we might need to execute the filter
expression after converting.

Best regards,
Hou zj