Re: [psycopg] speed concerns with executemany() - Mailing list psycopg

From Adrian Klaver
Subject Re: [psycopg] speed concerns with executemany()
Date
Msg-id 72b0ddc0-ee36-6a44-760f-7693f85e3304@aklaver.com
Whole thread Raw
In response to Re: [psycopg] speed concerns with executemany()  (Daniele Varrazzo <daniele.varrazzo@gmail.com>)
Responses Re: [psycopg] speed concerns with executemany()  (Christophe Pettus <xof@thebuild.com>)
List psycopg
On 12/23/2016 04:29 PM, Daniele Varrazzo wrote:
> On Sat, Dec 24, 2016 at 1:09 AM, Christophe Pettus <xof@thebuild.com> wrote:
>> Are you running with the transaction isolation level set to ISOLATION_LEVEL_AUTOCOMMIT?  If so, each of those
INSERTswill be in its own transaction, and thus will go through the COMMIT overhead.  That by itself wouldn't explain a
jumpthat large (in most environments), but it will definitely be *much* slower. 
>
> Why do you say this? Psycopg doesn't wrap statements in BEGIN/COMMIT
> when in autocommit mode. Are you referring about some implicit

I understood it did:

http://initd.org/psycopg/docs/usage.html#transactions-control

"It is possible to set the connection in autocommit mode: this way all
the commands executed will be immediately committed and no rollback is
possible. A few commands (e.g. CREATE DATABASE, VACUUM...) require to be
run outside any transaction: in order to be able to run these commands
from Psycopg, the connection must be in autocommit mode: you can use the
autocommit property (set_isolation_level() in older versions)."

> transaction created by the database?
>
> -- Daniele
>
>


--
Adrian Klaver
adrian.klaver@aklaver.com


psycopg by date:

Previous
From: Daniele Varrazzo
Date:
Subject: Re: [psycopg] speed concerns with executemany()
Next
From: Christophe Pettus
Date:
Subject: Re: [psycopg] speed concerns with executemany()