If You Subclass From Dict, Why Is __dict__ Always Empty?

I have a question regarding the subclassing behaviour in Python 2.7. If I subclass from the built-in dict type, it seems that __ dict __ is always empty. Where does Python save the

Solution 1:

__dict__ is where an object's attributes are stored. Dicts' items are not attributes, they are items. Most dicts have no data attributes.

BTW: subclassing dict is difficult to do properly, depending what you are trying to achieve. For example, you can override its __setitem__ method, but then update won't use it.

Solution 2:

A partial answer is that you're misunderstanding the purpose of __dict__. __dict__ is used to store attributes, not items, and it's present in most user-defined objects. Indeed, if you subclass dict in the appropriate way, __dict__will have values in it.

>>>classFoo(dict):...def__init__(self, *args, **kwargs):...super(Foo, self).__init__(*args, **kwargs)...        self.banana = 'banana'...        self['banana'] = 'not really a banana'...>>>f = Foo()>>>f.__dict__
{'banana': 'banana'}
>>>f.banana
'banana'
>>>f['banana']
'not really a banana'

Solution 3:

The dict class is implemented purely in C as a built-in. Its data storage is private to that implementation.

As a thought experiment, imagine if it put the name/value pairs into a Python dict, how would that dict store them? In another Python dict? And then, well, you get the idea!

Solution 4:

Because a dict object doesn't actually have __dict__ so the __dict__ you are referencing is the dict local to your object Foobar. Because Foobar has no attributes __dict__ is empty.

Solution 5:

__dict__ is for attributes:

>>>classD(dict):
    pass

>>>d=D()>>>d.__dict__
{}
>>>d.x=5>>>d.__dict__
{'x': 5}