From: | Karsten Hilbert <Karsten(dot)Hilbert(at)gmx(dot)net> |
---|---|
To: | psycopg(at)postgresql(dot)org |
Subject: | Re: Changing set_session implementation |
Date: | 2017-07-31 10:51:09 |
Message-ID: | 20170731105109.mjiimmdm3ixr7hyr@hermes.hilbert.loc |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | psycopg |
Hopefully I just hit summer vacation -- or else do I need to
provide more information to get an answer on the below ?
Thanks,
Karsten
On Thu, Jul 27, 2017 at 04:17:33PM +0200, Karsten Hilbert wrote:
> On Sat, Feb 04, 2017 at 04:33:13PM +0000, Daniele Varrazzo wrote:
>
> > I've changed the set_session/set_isolation_level/autocommit connection
> > methods and properties to use better the BEGIN parameters
> ...
> > Conversely:
> >
> > conn.set_session(readonly=True, autocommit=True)
>
> Does this mean that
>
> conn.set_isolation_level(ISOLATION_LEVEL_AUTOCOMMIT)
>
> is now a no-op as observed by the following log snippet or
> just diverted to set_session(autocommit=True) ?
>
>
> code:
>
> _log.debug('conn state now: %s', gmPG2.capture_conn_state(self.conn))
> _log.debug('setting isolation level to autocommit - should be redundant ?')
> self.conn.set_isolation_level(0)
> _log.debug('conn state now: %s', gmPG2.capture_conn_state(self.conn))
>
> log:
>
> 2017-07-27 16:03:01 DEBUG gm.bootstrapper MainThread (-1219281664) (./bootstrap_gm_db_system.py::__create_db() #810): conn state now: <connection object at 0xb64eca04; dsn: 'dbname=gnumed_v21 port=5432 user=postgres sslmode=prefer fallback_application_name=GNUmed client_encoding=utf8 application_name=postgres(at)template(dot)db', closed: 0>
> type: <class 'psycopg2.extras.DictConnection'>
> identity: 3058616836
> backend PID: 2864
> protocol version: 3
> encoding: UTF8
> isolation level (psyco): 3 (ISOLATION_LEVEL_SERIALIZABLE)
> readonly: False
> autocommit: True
> closed: 0
> connection status: 1 (STATUS_READY)
> transaction status: 0 (TRANSACTION_STATUS_IDLE)
> deferrable: None
> async: 0
> executing async op: False
> 2017-07-27 16:03:01 DEBUG gm.bootstrapper MainThread (-1219281664) (./bootstrap_gm_db_system.py::__create_db() #811): setting isolation level to autocommit - should be redundant ?
> 2017-07-27 16:03:01 DEBUG gm.bootstrapper MainThread (-1219281664) (./bootstrap_gm_db_system.py::__create_db() #813): conn state now: <connection object at 0xb64eca04; dsn: 'dbname=gnumed_v21 port=5432 user=postgres sslmode=prefer fallback_application_name=GNUmed client_encoding=utf8 application_name=postgres(at)template(dot)db', closed: 0>
> type: <class 'psycopg2.extras.DictConnection'>
> identity: 3058616836
> backend PID: 2864
> protocol version: 3
> encoding: UTF8
> isolation level (psyco): 3 (ISOLATION_LEVEL_SERIALIZABLE)
> readonly: False
> autocommit: True
> closed: 0
> connection status: 1 (STATUS_READY)
> transaction status: 0 (TRANSACTION_STATUS_IDLE)
> deferrable: None
> async: 0
> executing async op: False
>
>
> Note that it doesn't actually change the isolation level
> (which makes sense because autocommit mode is not an
> isolation level). Should conn.isolation_level show
> ISOLATION_LEVEL_DEFAULT in this case (meaning "use what the
> server/db is configured for") ? It doesn't seem to make a
> roundtrip to the server to _get_ the default because the
> database in question defaults to READ_COMMITED. The
> SERIALIZABLE results from an earlier call to
> .set_isolation_mode() in my code.
>
> (Also note that the connection had already been set to
> autocommit by "set_session(autocommit=True)" before.)
>
>
> Thanks for the clarification,
> Karsten
> --
> GPG key ID E4071346 @ eu.pool.sks-keyservers.net
> E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
>
>
> --
> Sent via psycopg mailing list (psycopg(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/psycopg
--
GPG key ID E4071346 @ eu.pool.sks-keyservers.net
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
From | Date | Subject | |
---|---|---|---|
Next Message | Stephen Quintero | 2017-08-02 04:20:44 | psycopg2: distinguishing connect failures through exception handling |
Previous Message | Karsten Hilbert | 2017-07-27 14:17:33 | Re: Changing set_session implementation |