On Sunday, February 9, 2014 3:14:50 PM UTC+2, Chris Angelico wrote:
> On Sun, Feb 9, 2014 at 11:47 PM, Asaf Las <r...@gmail.com> wrote:
>
Thanks
>
> Also, you're connecting and disconnecting repeatedly... oh, I see why
> it didn't work when I tried. You're also using two completely
> different database names: 'messageid.db' which is named in a constant
> and in the default argument, and 'example.db' which is what you
> actually use. Should have a single source of truth, otherwise you
> confuse the people who might otherwise be able to test your code :)
my apologies , it was deep night, when i got disappointed and forget to
update names so left as it is and did not check when posted :-)
> Anyway. This code is extremely inefficient:
>
> > conn = sqlite3.connect('example.db', 10.0, True, "EXCLUSIVE")
> > c = conn.cursor()
> > c.execute("SELECT val FROM msgid WHERE id = 0")
> > tint = int(c.fetchone()[0]) + 1
> > c.execute("UPDATE msgid SET val={0} WHERE id = 0".format(tint))
> > conn.commit()
> > conn.close()
>
> Much more common would be to retain a connection and repeatedly
> perform queries. Then you won't need to open it EXCLUSIVE, and you can
> simply query, update, and then commit (all your locks should be
> released at the commit). Do that, and you should see at least
> reasonable performance, plus everything should work correctly.
> ChrisA
i did it just to test sqlite3 behavior and actually test was related to
simulation of unique incremental sequence number/counter for
independently spawned tasks accessing counter in non deterministic manner.
/Asaf
--
https://mail.python.org/mailman/listinfo/python-list
