We are using sqlalchemy and postgres extensively, and have come across a
problem with deferred columns when an object is merged into a session with
load=False. Under these conditions, we find that the deferred columns are
not deferred - i.e. they loaded when any attribute is accessed, rather than
being deferred until explicitly accessed. We are using sqlalchemy 0.9.7 and
postgres 9.4.1.
Here's a simple example that illustrates the problem using the following
table:
class Port(Base):
''' Table port in schema comm. '''
__tablename__ = "port"
name = Column(String, primary_key=True)
port = deferred(Column(Integer, unique=True))
First we try querying Port on name:
my_port = session.query(Port).filter_by(name='test_thing').one()
my_port.name
This code produces the following query, showing that the port column is
correctly deferred:
2015-07-16 15:44:08.539 : sqlalchemy.engine.base.Engine :
base.py:912(_execute_context) : INFO : SELECT comm.port.name AS
comm_port_name
FROM comm.port
WHERE comm.port.name = %(name_1)s
Now I expire and expunge the my_port object, and then merge it back into
the session with load=False, and access name like this:
session.expire(my_port)
session.expunge(my_port)
new_port = session.merge(my_port, load=False)
new_port.name
Which produces the following query, showing that the port column has not
been deferred, and has instead been included in the query that fetches name:
2015-07-16 15:44:08.539 : sqlalchemy.engine.base.Engine :
base.py:912(_execute_context) : INFO : SELECT comm.port.port AS
comm_port_port, comm.port.name AS comm_port_name
FROM comm.port
WHERE comm.port.name = %(param_1)s
If I do not set load=False, the column is correctly deferred. I've also
tried creating a deferred object directly rather than expunging one from
the session and the problem still occurs, so it seems to be caused by the
load=False option.
Thanks,
Tom
--
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sqlalchemy.
For more options, visit https://groups.google.com/d/optout.
