I think my issue is that before the commit/rollback
Alice.id is None but after Alice.id is set to some integer. I understand the
object is transient but I wish that the primary key field was not modified in
this way because if I want to try to reinsert Alice in a new transaction I have
to write
defintions of transient, persistent:
https://docs.sqlalchemy.org/en/latest/orm/session_state_management.html#quickie-intro-to-object-states
note the objects have None for inspect(object).key, which is the
actual database identity.
from sqlalchemy import inspect
assert inspect(alice).key is None
On Tue, Sep 25, 2018 at 7:09 PM Doug Miller wrote:
>
> I've noticed that if a transaction fails midway through, objects that already
> have been inserted will have their primary keys set, even after rolling back.
>
> Partial snippet below:
>
> -
>
> class User(Base):
> id =
On Mon, Sep 24, 2018 at 12:22 PM YKdvd wrote:
>
> I have an ORM setup with a "departments_milestones" table ("dm",
> DepartmentMilestone objects), with "department_id" and "seqOrder" Fields.
> Each department has a few records in this table, ordered within the
> department by "seqOrder" (so
Huh… I use the -c option:
./bin/alembic -c proj-localhost.ini upgrade head
Jens
On Monday, September 24, 2018 at 11:21:42 PM UTC+10, Mike Bayer wrote:
>
>
> Looking at the source code, we already have "here":
>
> if self.config_file_name:
> here =
I've noticed that if a transaction fails midway through, objects that
already have been inserted will have their primary keys set, even after
rolling back.
Partial snippet below:
-
class User(Base):
id = Column(Integer, primary_key=True)
name = Column(String(80), unique=True,
got sql = "create table test( col timestamptz default ' now() at time
zone ''utc'' ' )
--
SQLAlchemy -
The Python SQL Toolkit and Object Relational Mapper
http://www.sqlalchemy.org/
To post example code, please provide an MCVE: Minimal, Complete, and Verifiable
Example. See
