Home > mailing lists

Re: [HACKERS] Bug in pg_dump --table and --exclude-table fordeclarative partition table handling. - Mailing list pgsql-hackers

From	Jeevan Ladhe
Subject	Re: [HACKERS] Bug in pg_dump --table and --exclude-table fordeclarative partition table handling.
Date	May 11, 2017 18:50:04
Msg-id	CAOgcT0PPxRsHMMhB0cZY=fC=O9y69tptH5PioHx64DPc9t14zA@mail.gmail.com Whole thread Raw
In response to	Re: [HACKERS] Bug in pg_dump --table and --exclude-table fordeclarative partition table handling. (Ashutosh Bapat <ashutosh.bapat@enterprisedb.com>)
Responses	Re: [HACKERS] Bug in pg_dump --table and --exclude-table for declarative partition table handling. (Tom Lane <tgl@sss.pgh.pa.us>)
List	pgsql-hackers

On Thu, May 11, 2017 at 4:47 PM, Ashutosh Bapat <ashutosh.bapat@enterprisedb.com> wrote:

We add PARTITION OF clause for a table which is partition, so if the
parent is not present while restoring, the restore is going to fail.

But, similarly for inheritance if we dump a child table, it's dump is broken as

of today. When we dump a child table we append "inherits(parenttab)" clause at

the end of the DDL. Later when we try to restore this table on a database not

having the parenttab, the restore fails.

So, I consider this as a bug.

Consider following example:

postgres=# create table tab1(a int, b int);

CREATE TABLE

postgres=# create table tab2(c int, d char) inherits(tab1);

CREATE TABLE

postgres=# \! pg_dump postgres -t tab2 > a.sql

postgres=# create database bkdb;

CREATE DATABASE

postgres=# \! psql bkdb < a.sql

SET

ERROR: relation "tab1" does not exist

ERROR: relation "tab2" does not exist

invalid command \.

postgres=#

Regards,

Jeevan Ladhe

From: Tom Lane
Date: 11 May 2017, 18:44:43
Subject: Re: [HACKERS] alter table..drop constraint pkey, left not null un-dropped

From: Simon Riggs
Date: 11 May 2017, 19:01:18
Subject: Re: [HACKERS] Time based lag tracking for logical replication