Yes it surprised me. But I get your point.
Thanks you for your response.
On Nov 26, 9:41 pm, Michael Bayer mike...@zzzcomputing.com wrote:
I wouldn't say its a bug since its intentional. But I'll grant the
intention is up for debate. I've always considered usage of execute() to
mean,
Hi,
We have a parent class/table (Notification/notifications) and a child
class/table(BusinessNotification/business_notifications).
Both of them have a version_id_col defined.
We are facing ConcurrentModificationException errors for the model at
the end of the mail.
The problem is because the
Hi,
After some more internal discussions we tried and realized that the version
col is not really required at the child table level.
The second commit in the sample code fails because the version value on the
parent row has already moved ahead.
Please confirm if the understanding/conclusion is
I might turn it on in 0.7 and add an extra flag autoflush_on_execute just so
people can turn it off if they don't like it. 0.7 is on track for beta
releases in december or january.
You can make yourself a quick Session subclass that wraps execute(), if you
wanted this behavior right now.
On Nov 29, 2010, at 8:28 AM, Naresh Khalasi wrote:
Hi,
After some more internal discussions we tried and realized that the version
col is not really required at the child table level.
The second commit in the sample code fails because the version value on the
parent row has already
we've got unicode round trips down very well for years now with plenty of
tests, so would need a specific series of steps to reproduce what you're doing
here. Note that the recommended connect string for MySQL + Mysqldb looks like
mysql://scott:ti...@localhost/test?charset=utf8use_unicode=0 .
Hi Michael
Thanks for your thoughts and comments to date.
I can replicate the problem with ease, so perhaps this will help;
# -*- coding: utf-8 -*-
e =
create_engine('mysql+mysqlconnector://user:passw...@127.0.0.1/testdb?use_unicode=0',
encoding='utf8', echo=False)
m = MetaData(e)
t =
Thanks Michael for the confirmation/explanation.
On Nov 29, 8:25 pm, Michael Bayer mike...@zzzcomputing.com wrote:
On Nov 29, 2010, at 8:28 AM, Naresh Khalasi wrote:
Hi,
After some more internal discussions we tried and realized that the version
col is not really required at the child