[sqlalchemy] Re: Help! More Class->Mapper confusion!
On Mar 11, 2009, at 6:05 PM, batraone wrote: > > Thanks. I don't have control over the instance creation. These pre- > created objects are handed down to me. I could create an adapter that > maps the original class to the one that has been mapped but this is > quite a bit of work as I have to manually copy over each of the source > class attributes to my mapped class. I could be clever and perhaps > look for common attribute names via the internal dictionary but this > too seems clunky. > > I would think this is a common problem (augmenting a class to dump > it's contents to a db but keep the original class untouched (sort of > like shelve - but w/o the > restore capability)). Is there a better way to handle this? Unfortunately the usage of an object relational mapper assumes that some degree of convention can be applied to the classes being mapped. For example, you can't map "old style" classes nor can you map most natively-backed classes (like cElementTree structures), instrumentation will fail if the class relies upon direct __dict__ access, and there's probably many other examples of limitations here. Applying instrumentation to a class to which you don't have control over is generally a bad idea no matter what persistence approach you're using, as even if you do get it working, changes to the implementation of those objects will break your application. So in this case you definitely should be marshalling these external objects immediately into an object structure which you have control over.You also should most certainly be using standard Python idioms like dir(object), getattr() and possibly __dict__ access in order to marshall the state of the incoming objects into your own model using a reflective approach rather than hardcoding every attribute. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Help! More Class->Mapper confusion!
Thanks. I don't have control over the instance creation. These pre- created objects are handed down to me. I could create an adapter that maps the original class to the one that has been mapped but this is quite a bit of work as I have to manually copy over each of the source class attributes to my mapped class. I could be clever and perhaps look for common attribute names via the internal dictionary but this too seems clunky. I would think this is a common problem (augmenting a class to dump it's contents to a db but keep the original class untouched (sort of like shelve - but w/o the restore capability)). Is there a better way to handle this? Thanks, Raj On Mar 11, 2:48 pm, "Michael Bayer" wrote: > batraone wrote: > > def map(self, myfoo): > > 'Creates the map. ' > > > orm.mapper(Foo, self.t_foo, column_prefix = '_') > > > # Add foo to the database > > print dir(myfoo) # where did _name go?, _phone is there! > > > mf2 = Foo() # ok, let's create a new one. > > print dir(mf2) # same problem > > > if __name__ == '__main__': > > f = Foo() > > fs = FooStore() > > fs.map(f) > > The column_prefix wasn't being honored when the mapper checked for > existing names, this is fixed in trunk r5839. > > But also, don't create instances of the object before the class is mapped. > In particular, it's bad form to create tables and mappers inside of class > methods. Create class-level constructs like tables and mappers at the > module level, in the same scope in which you create your classes. > > > > > mf2 = Foo.Foo() # ok, let's create a new one. > > # AttributeError: 'Foo' object has no attribute > > '_sa_instance_state' > > myfoo._phone = '555-1212' # > > > if __name__ == '__main__': > > orm.clear_mappers() > > f = Foo.Foo() > > fs = FooStore() > > fs.map(f) > > don't create instances of the object before the class is mapped.- Hide quoted > text - > > - Show quoted text - --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
[sqlalchemy] Re: Help! More Class->Mapper confusion!
batraone wrote: > def map(self, myfoo): > 'Creates the map. ' > > orm.mapper(Foo, self.t_foo, column_prefix = '_') > > # Add foo to the database > print dir(myfoo) # where did _name go?, _phone is there! > > mf2 = Foo() # ok, let's create a new one. > print dir(mf2) # same problem > > if __name__ == '__main__': > f = Foo() > fs = FooStore() > fs.map(f) The column_prefix wasn't being honored when the mapper checked for existing names, this is fixed in trunk r5839. But also, don't create instances of the object before the class is mapped. In particular, it's bad form to create tables and mappers inside of class methods. Create class-level constructs like tables and mappers at the module level, in the same scope in which you create your classes. > > mf2 = Foo.Foo() # ok, let's create a new one. > # AttributeError: 'Foo' object has no attribute > '_sa_instance_state' > myfoo._phone = '555-1212' # > > if __name__ == '__main__': > orm.clear_mappers() > f = Foo.Foo() > fs = FooStore() > fs.map(f) don't create instances of the object before the class is mapped. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~--~~~~--~~--~--~---
