I’m trying to figure out how to map against a simple read-only property and have that property fire when I save to the database.
A contrived example should make this more clear. First, a simple table:
meta = MetaData()
foo_table = Table('foo', meta,
Column('id', String(3), primary_key=True),
Column('description', String(64), nullable=False),
Column('calculated_value', Integer, nullable=False),
)
What I want to do is set up a class with a read-only property that will insert into the calculated_value column for me when I call session.commit()…
import datetime
def Foo(object):
def __init__(self, id, description):
self.id = id
self.description = description
@property
def calculated_value(self):
self._calculated_value = datetime.datetime.now().second + 10
return self._calculated_value
According to the sqlalchemy docs, I think I am supposed to map this like so:
mapper(Foo, foo_table, properties = {
'calculated_value' : synonym('_calculated_value', map_column=True)
})
The problem with this is that _calculated_value is None until you access the calculated_value property. It appears that SQLAlchemy is not calling the property on insertion into the database, so I’m getting a None value instead. What is the correct way to map this so that the result of the “calculated_value” property is inserted into the foo table’s “calculated_value” column?
OK – I am editing this post in case someone else has the same question. What I ended up doing was using a MapperExtension. Let me give you a better example along with usage of the extension:
class UpdatePropertiesExtension(MapperExtension):
def __init__(self, properties):
self.properties = properties
def _update_properties(self, instance):
# We simply need to access our read only property one time before it gets
# inserted into the database.
for property in self.properties:
getattr(instance, property)
def before_insert(self, mapper, connection, instance):
self._update_properties(instance)
def before_update(self, mapper, connection, instance):
self._update_properties(instance)
And this is how you use this. Lets say you have a class with several read only properties that must fire before insertion into the database. I am assuming here that for each one of these read only properties, you have a corresponding column in the database that you want populated with the value of the property. You are still going to set up a synonym for each property, but you use the mapper extension above when you map the object:
class Foo(object):
def __init__(self, id, description):
self.id = id
self.description = description
self.items = []
self.some_other_items = []
@property
def item_sum(self):
self._item_sum = 0
for item in self.items:
self._item_sum += item.some_value
return self._item_sum
@property
def some_other_property(self):
self._some_other_property = 0
.... code to generate _some_other_property on the fly....
return self._some_other_property
mapper(Foo, metadata,
extension = UpdatePropertiesExtension(['item_sum', 'some_other_property']),
properties = {
'item_sum' : synonym('_item_sum', map_column=True),
'some_other_property' : synonym('_some_other_property', map_column = True)
})
I’m not sure it’s possible to achieve what you want using sqlalchemy.orm.synonym. Propably not given the fact how sqlalchemy keeps track of which instances are dirty and need to be updated during flush.
But there is other way how you can get this functionality – SessionExtensions (notice the engine_string variable at the top that needs to be filled):
More on SessionExtensions: sqlalchemy.orm.interfaces.SessionExtension.