Given:
In [37]: class A:
....: f = 1
....:
In [38]: class B(A):
....: pass
....:
In [39]: getattr(B, 'f')
Out[39]: 1
Okay, that either calls super or crawls the mro?
In [40]: getattr(A, 'f')
Out[40]: 1
This is expected.
In [41]: object.__getattribute__(A, 'f')
Out[41]: 1
In [42]: object.__getattribute__(B, 'f')
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-42-de76df798d1d> in <module>()
----> 1 object.__getattribute__(B, 'f')
AttributeError: 'type' object has no attribute 'f'
What is getattribute not doing that getattr does?
In [43]: type.__getattribute__(B, 'f')
Out[43]: 1
What?! type.__getattribute__
calls super but object
's version doesn't?
In [44]: type.__getattribute__(A, 'f')
Out[44]: 1
__getattribute__This method should return the (computed) attribute value or raise an AttributeError exception. In order to avoid infinite recursion in this method, its implementation should always call the base class method with the same name to access any attributes it needs, for example, object.
Python getattr() function is used to get the value of an object's attribute and if no attribute of that object is found, default value is returned. Basically, returning the default value is the main reason why you may need to use Python getattr() function.
A special attribute of every module is __dict__. This is the dictionary containing the module's symbol table. A dictionary or other mapping object used to store an object's (writable) attributes.
Python setattr() and getattr() goes hand-in-hand. As we have already seen what getattr() does; The setattr() function is used to assign a new value to an object/instance attribute. Syntax. Use a different Browser.
You are operating directly on classes. object.__getattribute__
is only used on instances of A
and B
instead. That's because special methods are looked up on the type; for instances the type is the class.
For classes then, the type is.. type
:
>>> class A:
... f = 1
...
>>> class B(A):
... pass
...
>>> type(B)
<class 'type'>
so type.__getattribute__
is used:
>>> type.__getattribute__(B, 'f')
1
and object.__getattribute__
works fine on instances:
>>> object.__getattribute__(B(), 'f')
1
For instances attributes are looked up first on the class (in the case of data descriptors), then on the instance, then if the instance doesn't have the attribute, the class hierarchy is searched in MRO order. This is the job of object.__getattribute__
. So object.__getattribute__
looks at the first argument (e.g. self
, the instance object) for the attribute, and at objects in type(self).__mro__
.
For classes, attributes are looked up on the class itself and all its bases; type.__getattribute__
looks directly at self.__mro__
for these; self
being a class object here.
If you use object.__getattribute__
for classes then, there is no f
attribute on B
directly, and no f
anywhere in type(B).__mro__
. If you use type.__getattribute__
, A
is a member of B.__mro__
so f
is found there:
>>> type(B).__mro__
(<class 'type'>, <class 'object'>)
>>> B.__mro__
(<class '__main__.B'>, <class '__main__.A'>, <class 'object'>)
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With