Column rename in an extension update script
Hi all,
I am coding an update script for an extension. And I am in trouble when
trying to rename a column of an existing table.
Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.
I wrote a simple test case to show this. Here is the shell script that
can be easily adapted.
# issue in postgres extension when trying to access a column that has
been renamed inside an extension update script
#
export EXTDIR="/tmp"
export PGDIR="/usr/local/pg962/share/postgresql/extension"
export PGHOST=localhost
export PGPORT=5496
export PGDATABASE='postgres'
echo "create files for the extension"
echo "------------------------------"
cat >$EXTDIR/myextension.control <<*END*
default_version = '1'
directory = '$EXTDIR'
*END*
sudo ln -s $EXTDIR/myextension.control $PGDIR/myextension.control
cat >$EXTDIR/myextension--1.sql <<*END*
CREATE TABLE mytable (col_old INT);
*END*
cat >$EXTDIR/myextension--1--2.sql <<*END*
ALTER TABLE mytable RENAME col_old TO col_new;
UPDATE mytable SET col_new = 0;
*END*
echo "psql: run the test ==> FAILS"
echo "----------------------------"
psql -a <<*END*
select version();
CREATE EXTENSION myextension VERSION '1';
ALTER EXTENSION myextension UPDATE TO '2';
DROP EXTENSION IF EXISTS myextension;
*END*
echo "psql: similar statements outside extension ==> WORKS"
echo "----------------------------------------------------"
psql -a <<*END*
CREATE TABLE mytable (col_old INT);
BEGIN;
ALTER TABLE mytable RENAME col_old TO col_new;
UPDATE mytable SET col_new = 0;
COMMIT;
DROP TABLE IF EXISTS mytable;
*END*
sudo rm $PGDIR/myextension.control
rm $EXTDIR/myextension*
And here is the result:
create files for the extension
------------------------------
psql: run the test ==> FAILS
----------------------------
select version();
version
-----------------------------------------------------------------------------------------------------------------
PostgreSQL 9.6.2 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu
5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609, 64-bit
(1 row)
CREATE EXTENSION myextension VERSION '1';
CREATE EXTENSION
ALTER EXTENSION myextension UPDATE TO '2';
ERROR: column "col_new" of relation "mytable" does not exist
DROP EXTENSION IF EXISTS myextension;
DROP EXTENSION
psql: similar statements outside extension ==> WORKS
----------------------------------------------------
CREATE TABLE mytable (col_old INT);
CREATE TABLE
BEGIN;
BEGIN
ALTER TABLE mytable RENAME col_old TO col_new;
ALTER TABLE
UPDATE mytable SET col_new = 0;
UPDATE 0
COMMIT;
COMMIT
DROP TABLE IF EXISTS mytable;
DROP TABLE
As you can see:
- the error message is "ERROR: column "col_new" of relation "mytable"
does not exist", while the ALTER TABLE statement doesn't return any error,
- the same statements in a simple psql script works fine,
- I reproduce this with all supported postgres versions.
As a workaround, I perform the UPDATE statement before the ALTER TABLE
operation, using of course the old column name.
I probably do something wrong. But I can't see what.
Thanks by advance for any piece of advise.
Best regards. Philippe Beaudoin.
moving this thread to -hackers, since this looks like a bug.
On 01/05/2017 08:54, Philippe BEAUDOIN wrote:
Hi all,
I am coding an update script for an extension. And I am in trouble when
trying to rename a column of an existing table.Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.
I can reproduce this issue.
It looks like this is due to missing cache invalidation between the
ALTER TABLE execution and next query planning (failure happens during
pg_analyze_and_rewrite()).
AFAICT, CommandCounterIncrement() is responsible for handling
invalidation messages, but in execute_sql_string() this function is
called before executing a Stmt instead of after. Moving the
CommandCounterIncrement() just before the PopActiveSnapshot() call (and
removing the final one) fixes this issue for me, but I have no idea if
this is the right fix.
--
Julien Rouhaud
http://dalibo.com - http://dalibo.org
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Julien Rouhaud <julien.rouhaud@dalibo.com> writes:
moving this thread to -hackers, since this looks like a bug.
On 01/05/2017 08:54, Philippe BEAUDOIN wrote:I am coding an update script for an extension. And I am in trouble when
trying to rename a column of an existing table.Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.
I can reproduce this issue.
It looks like this is due to missing cache invalidation between the
ALTER TABLE execution and next query planning (failure happens during
pg_analyze_and_rewrite()).
Yes. Kind of surprising nobody noticed this before --- you'd certainly
think that extension scripts would have lots of cases of statements
depending on DDL done just before them. I think it must have been masked
by the fact that many DDL commands *end* with CommandCounterIncrement,
or at least have one pretty close to the end. But not renameatt().
AFAICT, CommandCounterIncrement() is responsible for handling
invalidation messages, but in execute_sql_string() this function is
called before executing a Stmt instead of after. Moving the
CommandCounterIncrement() just before the PopActiveSnapshot() call (and
removing the final one) fixes this issue for me, but I have no idea if
this is the right fix.
I'm inclined to add one before the parse step, rather than removing any.
The sequence of bump-the-command-counter-then-capture-a-snapshot is
pretty well established in places like spi.c, so I don't want to change
that usage. Also, I think part of the point here was to ensure that
any DDL done *before* reaching execute_sql_string() is visible to the
first command. Also note the assumption in ApplyExtensionUpdates that
execute_sql_string will do at least one CommandCounterIncrement, even
if the script is empty. A CCI that has nothing to do is quite cheap,
and we're not that worried about performance here anyway IMO, so I'd
just as soon err on the side of having more than enough CCIs.
regards, tom lane
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
On 04/30/2017 11:54 PM, Philippe BEAUDOIN wrote:
Hi all,
I am coding an update script for an extension. And I am in trouble when
trying to rename a column of an existing table.Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.
From the error it looks to me like the statements are each run in a
separate session and the UPDATE is not seeing the ALTER TABLE. A quick
search of the source indicates this is handled in extension.c:
In particular execute_sql_string line 684. I do not understand C well
enough to figure out if the above is actually creating separate sessions
or not. Maybe you understand it or someone else can chime in.
# issue in postgres extension when trying to access a column that has
been renamed inside an extension update script
#
export EXTDIR="/tmp"
export PGDIR="/usr/local/pg962/share/postgresql/extension"
export PGHOST=localhost
export PGPORT=5496
export PGDATABASE='postgres'echo "create files for the extension"
echo "------------------------------"
cat >$EXTDIR/myextension.control <<*END*
default_version = '1'
directory = '$EXTDIR'
*END*
sudo ln -s $EXTDIR/myextension.control $PGDIR/myextension.controlcat >$EXTDIR/myextension--1.sql <<*END*
CREATE TABLE mytable (col_old INT);
*END*cat >$EXTDIR/myextension--1--2.sql <<*END*
ALTER TABLE mytable RENAME col_old TO col_new;
UPDATE mytable SET col_new = 0;
*END*echo "psql: run the test ==> FAILS"
echo "----------------------------"
psql -a <<*END*
select version();
CREATE EXTENSION myextension VERSION '1';
ALTER EXTENSION myextension UPDATE TO '2';
DROP EXTENSION IF EXISTS myextension;
*END*echo "psql: similar statements outside extension ==> WORKS"
echo "----------------------------------------------------"
psql -a <<*END*
CREATE TABLE mytable (col_old INT);
BEGIN;
ALTER TABLE mytable RENAME col_old TO col_new;
UPDATE mytable SET col_new = 0;
COMMIT;
DROP TABLE IF EXISTS mytable;
*END*sudo rm $PGDIR/myextension.control
rm $EXTDIR/myextension*And here is the result:
create files for the extension
------------------------------
psql: run the test ==> FAILS
----------------------------
select version();version
-----------------------------------------------------------------------------------------------------------------
PostgreSQL 9.6.2 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu
5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609, 64-bit
(1 row)CREATE EXTENSION myextension VERSION '1';
CREATE EXTENSION
ALTER EXTENSION myextension UPDATE TO '2';
ERROR: column "col_new" of relation "mytable" does not exist
DROP EXTENSION IF EXISTS myextension;
DROP EXTENSION
psql: similar statements outside extension ==> WORKS
----------------------------------------------------
CREATE TABLE mytable (col_old INT);
CREATE TABLE
BEGIN;
BEGIN
ALTER TABLE mytable RENAME col_old TO col_new;
ALTER TABLE
UPDATE mytable SET col_new = 0;
UPDATE 0
COMMIT;
COMMIT
DROP TABLE IF EXISTS mytable;
DROP TABLEAs you can see:
- the error message is "ERROR: column "col_new" of relation "mytable"
does not exist", while the ALTER TABLE statement doesn't return any error,- the same statements in a simple psql script works fine,
- I reproduce this with all supported postgres versions.
As a workaround, I perform the UPDATE statement before the ALTER TABLE
operation, using of course the old column name.I probably do something wrong. But I can't see what.
Thanks by advance for any piece of advise.
Best regards. Philippe Beaudoin.
--
Adrian Klaver
adrian.klaver@aklaver.com
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Adrian Klaver <adrian.klaver@aklaver.com> writes:
On 04/30/2017 11:54 PM, Philippe BEAUDOIN wrote:
Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.
From the error it looks to me like the statements are each run in a
separate session and the UPDATE is not seeing the ALTER TABLE.
No, it's in the same session; the problem is the lack of a
CommandCounterIncrement call between the ALTER's update and the parsing
of the next statement. That means the update isn't visible yet,
even in its own session. See the fix here:
regards, tom lane
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Le 03/05/2017 � 19:29, Tom Lane a �crit :
Adrian Klaver <adrian.klaver@aklaver.com> writes:
On 04/30/2017 11:54 PM, Philippe BEAUDOIN wrote:
Just after the ALTER TABLE statement, I want to access this table. But
at this time, the altered column is not visible with its new name.From the error it looks to me like the statements are each run in a
separate session and the UPDATE is not seeing the ALTER TABLE.No, it's in the same session; the problem is the lack of a
CommandCounterIncrement call between the ALTER's update and the parsing
of the next statement. That means the update isn't visible yet,
even in its own session. See the fix here:regards, tom lane
Thanks Tom for the fix. And thanks to Julien and Adrian too, for the
time spent on this issue.
Regards.
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general