Do not define the mapping for the columns which you want to load on demand. Then configure those as described in the Deferred Column Loading using mapper
object. Modified code here:
from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class SomeClass(Base):
__tablename__ = 'some_table'
id = Column(Integer, primary_key=True)
name = Column(String(50))
#big_name = Column(String(500))
SomeClass.__table__.append_column(Column('big_name', String(500)))
SomeClass.__mapper__.add_property('big_name', deferred(SomeClass.__table__.c.big_name))
Running this test code:
c = session.query(SomeClass).first()
# here SQL is loading all configured properties, but big_name
print "c: ", c
# only here another SQL request is made to load the property
print "big_name: ", c.big_name
produces log extract:
... INFO sqlalchemy.engine.base.Engine.0x...77d0 SELECT some_table.id AS some_table_id, some_table.name AS some_table_name
FROM some_table
LIMIT 1 OFFSET 0
... INFO sqlalchemy.engine.base.Engine.0x...77d0 SELECT some_table.big_name AS some_table_big_name
FROM some_table
WHERE some_table.id = ?