Python的super（）如何与多重继承一起使用？

I’m pretty much new in Python object oriented programming and I have trouble understanding the super() function (new style classes) especially when it comes to multiple inheritance.

For example if you have something like:

class First(object):
    def __init__(self):
        print "first"

class Second(object):
    def __init__(self):
        print "second"

class Third(First, Second):
    def __init__(self):
        super(Third, self).__init__()
        print "that's it"

What I don’t get is: will the Third() class inherit both constructor methods? If yes, then which one will be run with super() and why?

And what if you want to run the other one? I know it has something to do with Python method resolution order (MRO).

This is detailed with a reasonable amount of detail by Guido himself in his blog post Method Resolution Order (including two earlier attempts).

In your example, Third() will call First.__init__. Python looks for each attribute in the class’s parents as they are listed left to right. In this case, we are looking for __init__. So, if you define

class Third(First, Second):
    ...

Python will start by looking at First, and, if First doesn’t have the attribute, then it will look at Second.

This situation becomes more complex when inheritance starts crossing paths (for example if First inherited from Second). Read the link above for more details, but, in a nutshell, Python will try to maintain the order in which each class appears on the inheritance list, starting with the child class itself.

So, for instance, if you had:

class First(object):
    def __init__(self):
        print "first"

class Second(First):
    def __init__(self):
        print "second"

class Third(First):
    def __init__(self):
        print "third"

class Fourth(Second, Third):
    def __init__(self):
        super(Fourth, self).__init__()
        print "that's it"

the MRO would be [Fourth, Second, Third, First].

By the way: if Python cannot find a coherent method resolution order, it’ll raise an exception, instead of falling back to behavior which might surprise the user.

Edited to add an example of an ambiguous MRO:

class First(object):
    def __init__(self):
        print "first"

class Second(First):
    def __init__(self):
        print "second"

class Third(First, Second):
    def __init__(self):
        print "third"

Should Third‘s MRO be [First, Second] or [Second, First]? There’s no obvious expectation, and Python will raise an error:

TypeError: Error when calling the metaclass bases
    Cannot create a consistent method resolution order (MRO) for bases Second, First

Edit: I see several people arguing that the examples above lack super() calls, so let me explain: The point of the examples is to show how the MRO is constructed. They are not intended to print “first\nsecond\third” or whatever. You can – and should, of course, play around with the example, add super() calls, see what happens, and gain a deeper understanding of Python’s inheritance model. But my goal here is to keep it simple and show how the MRO is built. And it is built as I explained:

>>> Fourth.__mro__
(<class '__main__.Fourth'>,
 <class '__main__.Second'>, <class '__main__.Third'>,
 <class '__main__.First'>,
 <type 'object'>)

Your code, and the other answers, are all buggy. They are missing the super() calls in the first two classes that are required for co-operative subclassing to work.

Here is a fixed version of the code:

class First(object):
    def __init__(self):
        super(First, self).__init__()
        print("first")

class Second(object):
    def __init__(self):
        super(Second, self).__init__()
        print("second")

class Third(First, Second):
    def __init__(self):
        super(Third, self).__init__()
        print("third")

The super() call finds the next method in the MRO at each step, which is why First and Second have to have it too, otherwise execution stops at the end of Second.__init__().

This is what I get:

>>> Third()
second
first
third

I wanted to elaborate the answer by lifeless a bit because when I started reading about how to use super() in a multiple inheritance hierarchy in Python, I did’t get it immediately.

What you need to understand is that super(MyClass, self).__init__() provides the next __init__ method according to the used Method Resolution Ordering (MRO) algorithm in the context of the complete inheritance hierarchy.

This last part is crucial to understand. Let’s consider the example again:

#!/usr/bin/env python2

class First(object):
  def __init__(self):
    print "First(): entering"
    super(First, self).__init__()
    print "First(): exiting"

class Second(object):
  def __init__(self):
    print "Second(): entering"
    super(Second, self).__init__()
    print "Second(): exiting"

class Third(First, Second):
  def __init__(self):
    print "Third(): entering"
    super(Third, self).__init__()
    print "Third(): exiting"

According to this article about Method Resolution Order by Guido van Rossum, the order to resolve __init__ is calculated (before Python 2.3) using a “depth-first left-to-right traversal” :

Third --> First --> object --> Second --> object

After removing all duplicates, except for the last one, we get :

Third --> First --> Second --> object

So, lets follow what happens when we instantiate an instance of the Third class, e.g. x = Third().

1. According to MRO Third.__init__ executes.
   • prints Third(): entering
   • then super(Third, self).__init__() executes and MRO returns First.__init__ which is called.
2. First.__init__ executes.
   • prints First(): entering
   • then super(First, self).__init__() executes and MRO returns Second.__init__ which is called.
3. Second.__init__ executes.
   • prints Second(): entering
   • then super(Second, self).__init__() executes and MRO returns object.__init__ which is called.
4. object.__init__ executes (no print statements in the code there)
5. execution goes back to Second.__init__ which then prints Second(): exiting
6. execution goes back to First.__init__ which then prints First(): exiting
7. execution goes back to Third.__init__ which then prints Third(): exiting

This details out why instantiating Third() results in to :

Third(): entering
First(): entering
Second(): entering
Second(): exiting
First(): exiting
Third(): exiting

The MRO algorithm has been improved from Python 2.3 onwards to work well in complex cases, but I guess that using the “depth-first left-to-right traversal” + “removing duplicates expect for the last” still works in most cases (please comment if this is not the case). Be sure to read the blog post by Guido!

This is known as the Diamond Problem, the page has an entry on Python, but in short, Python will call the superclass’s methods from left to right.

This is to how I solved to issue of having multiple inheritance with different variables for initialization and having multiple MixIns with the same function call. I had to explicitly add variables to passed **kwargs and add a MixIn interface to be an endpoint for super calls.

Here A is an extendable base class and B and C are MixIn classes both who provide function f. A and B both expect parameter v in their __init__ and C expects w. The function f takes one parameter y. Q inherits from all three classes. MixInF is the mixin interface for B and C.

• IPython NoteBook Of This Code
• Github Repo with code example

class A(object):
    def __init__(self, v, *args, **kwargs):
        print "A:init:v[{0}]".format(v)
        kwargs['v']=v
        super(A, self).__init__(*args, **kwargs)
        self.v = v


class MixInF(object):
    def __init__(self, *args, **kwargs):
        print "IObject:init"
    def f(self, y):
        print "IObject:y[{0}]".format(y)


class B(MixInF):
    def __init__(self, v, *args, **kwargs):
        print "B:init:v[{0}]".format(v)
        kwargs['v']=v
        super(B, self).__init__(*args, **kwargs)
        self.v = v
    def f(self, y):
        print "B:f:v[{0}]:y[{1}]".format(self.v, y)
        super(B, self).f(y)


class C(MixInF):
    def __init__(self, w, *args, **kwargs):
        print "C:init:w[{0}]".format(w)
        kwargs['w']=w
        super(C, self).__init__(*args, **kwargs)
        self.w = w
    def f(self, y):
        print "C:f:w[{0}]:y[{1}]".format(self.w, y)
        super(C, self).f(y)


class Q(C,B,A):
    def __init__(self, v, w):
        super(Q, self).__init__(v=v, w=w)
    def f(self, y):
        print "Q:f:y[{0}]".format(y)
        super(Q, self).f(y)

I understand this doesn’t directly answer the super() question, but I feel it’s relevant enough to share.

There is also a way to directly call each inherited class:

class First(object):
    def __init__(self):
        print '1'

class Second(object):
    def __init__(self):
        print '2'

class Third(First, Second):
    def __init__(self):
        Second.__init__(self)

Just note that if you do it this way, you’ll have to call each manually as I’m pretty sure First‘s __init__() won’t be called.

Overall

Assuming everything descends from object (you are on your own if it doesn’t), Python computes a method resolution order (MRO) based on your class inheritance tree. The MRO satisfies 3 properties:

• Children of a class come before their parents
• Left parents come before right parents
• A class only appears once in the MRO

If no such ordering exists, Python errors. The inner workings of this is a C3 Linerization of the classes ancestry. Read all about it here: https://www.python.org/download/releases/2.3/mro/

Thus, in both of the examples below, it is:

1. Child
2. Left
3. Right
4. Parent

When a method is called, the first occurrence of that method in the MRO is the one that is called. Any class that doesn’t implement that method is skipped. Any call to super within that method will call the next occurrence of that method in the MRO. Consequently, it matters both what order you place classes in inheritance, and where you put the calls to super in the methods.

With super first in each method

class Parent(object):
    def __init__(self):
        super(Parent, self).__init__()
        print "parent"

class Left(Parent):
    def __init__(self):
        super(Left, self).__init__()
        print "left"

class Right(Parent):
    def __init__(self):
        super(Right, self).__init__()
        print "right"

class Child(Left, Right):
    def __init__(self):
        super(Child, self).__init__()
        print "child"

Child() Outputs:

parent
right
left
child

With super last in each method

class Parent(object):
    def __init__(self):
        print "parent"
        super(Parent, self).__init__()

class Left(Parent):
    def __init__(self):
        print "left"
        super(Left, self).__init__()

class Right(Parent):
    def __init__(self):
        print "right"
        super(Right, self).__init__()

class Child(Left, Right):
    def __init__(self):
        print "child"
        super(Child, self).__init__()

Child() Outputs:

child
left
right
parent

About @calfzhou’s comment, you can use, as usually, **kwargs:

Online running example

class A(object):
  def __init__(self, a, *args, **kwargs):
    print("A", a)

class B(A):
  def __init__(self, b, *args, **kwargs):
    super(B, self).__init__(*args, **kwargs)
    print("B", b)

class A1(A):
  def __init__(self, a1, *args, **kwargs):
    super(A1, self).__init__(*args, **kwargs)
    print("A1", a1)

class B1(A1, B):
  def __init__(self, b1, *args, **kwargs):
    super(B1, self).__init__(*args, **kwargs)
    print("B1", b1)


B1(a1=6, b1=5, b="hello", a=None)

Result:

A None
B hello
A1 6
B1 5

You can also use them positionally:

B1(5, 6, b="hello", a=None)

but you have to remember the MRO, it’s really confusing.

I can be a little annoying, but I noticed that people forgot every time to use *args and **kwargs when they override a method, while it’s one of few really useful and sane use of these ‘magic variables’.

Another not yet covered point is passing parameters for initialization of classes. Since the destination of super depends on the subclass the only good way to pass parameters is packing them all together. Then be careful to not have the same parameter name with different meanings.

Example:

class A(object):
    def __init__(self, **kwargs):
        print('A.__init__')
        super().__init__()

class B(A):
    def __init__(self, **kwargs):
        print('B.__init__ {}'.format(kwargs['x']))
        super().__init__(**kwargs)


class C(A):
    def __init__(self, **kwargs):
        print('C.__init__ with {}, {}'.format(kwargs['a'], kwargs['b']))
        super().__init__(**kwargs)


class D(B, C): # MRO=D, B, C, A
    def __init__(self):
        print('D.__init__')
        super().__init__(a=1, b=2, x=3)

print(D.mro())
D()

gives:

[<class '__main__.D'>, <class '__main__.B'>, <class '__main__.C'>, <class '__main__.A'>, <class 'object'>]
D.__init__
B.__init__ 3
C.__init__ with 1, 2
A.__init__

Calling the super class __init__ directly to more direct assignment of parameters is tempting but fails if there is any super call in a super class and/or the MRO is changed and class A may be called multiple times, depending on the implementation.

To conclude: cooperative inheritance and super and specific parameters for initialization aren’t working together very well.

class First(object):
  def __init__(self, a):
    print "first", a
    super(First, self).__init__(20)

class Second(object):
  def __init__(self, a):
    print "second", a
    super(Second, self).__init__()

class Third(First, Second):
  def __init__(self):
    super(Third, self).__init__(10)
    print "that's it"

t = Third()

Output is

first 10
second 20
that's it

Call to Third() locates the init defined in Third. And call to super in that routine invokes init defined in First. MRO=[First, Second]. Now call to super in init defined in First will continue searching MRO and find init defined in Second, and any call to super will hit the default object init. I hope this example clarifies the concept.

If you don’t call super from First. The chain stops and you will get the following output.

first 10
that's it

In learningpythonthehardway I learn something called super() an in-built function if not mistaken. Calling super() function can help the inheritance to pass through the parent and ‘siblings’ and help you to see clearer. I am still a beginner but I love to share my experience on using this super() in python2.7.

If you have read through the comments in this page, you will hear of Method Resolution Order (MRO), the method being the function you wrote, MRO will be using Depth-First-Left-to-Right scheme to search and run. You can do more research on that.

By adding super() function

super(First, self).__init__() #example for class First.

You can connect multiple instances and ‘families’ with super(), by adding in each and everyone in them. And it will execute the methods, go through them and make sure you didn’t miss out! However, adding them before or after does make a difference you will know if you have done the learningpythonthehardway exercise 44. Let the fun begins!!

Taking example below, you can copy & paste and try run it:

class First(object):
    def __init__(self):

        print("first")

class Second(First):
    def __init__(self):
        print("second (before)")
        super(Second, self).__init__()
        print("second (after)")

class Third(First):
    def __init__(self):
        print("third (before)")
        super(Third, self).__init__()
        print("third (after)")


class Fourth(First):
    def __init__(self):
        print("fourth (before)")
        super(Fourth, self).__init__()
        print("fourth (after)")


class Fifth(Second, Third, Fourth):
    def __init__(self):
        print("fifth (before)")
        super(Fifth, self).__init__()
        print("fifth (after)")

Fifth()

How does it run? The instance of fifth() will goes like this. Each step goes from class to class where the super function added.

1.) print("fifth (before)")
2.) super()>[Second, Third, Fourth] (Left to right)
3.) print("second (before)")
4.) super()> First (First is the Parent which inherit from object)

The parent was found and it will go continue to Third and Fourth!!

5.) print("third (before)")
6.) super()> First (Parent class)
7.) print ("Fourth (before)")
8.) super()> First (Parent class)

Now all the classes with super() have been accessed! The parent class has been found and executed and now it continues to unbox the function in the inheritances to finished the codes.

9.) print("first") (Parent)
10.) print ("Fourth (after)") (Class Fourth un-box)
11.) print("third (after)") (Class Third un-box)
12.) print("second (after)") (Class Second un-box)
13.) print("fifth (after)") (Class Fifth un-box)
14.) Fifth() executed

The outcome of the program above:

fifth (before)
second (before
third (before)
fourth (before)
first
fourth (after)
third (after)
second (after)
fifth (after)

For me by adding super() allows me to see clearer on how python would execute my coding and make sure the inheritance can access the method I intended.

I would like to add to what @Visionscaper says at the top:

Third --> First --> object --> Second --> object

In this case the interpreter doesnt filter out the object class because its duplicated, rather its because Second appears in a head position and doesnt appear in the tail position in a hierarchy subset. While object only appears in tail positions and is not considered a strong position in C3 algorithm to determine priority.

The linearisation(mro) of a class C, L(C), is the

• the Class C
• plus the merge of
  • linearisation of its parents P1, P2, .. = L(P1, P2, …) and
  • the list of its parents P1, P2, ..

Linearised Merge is done by selecting the common classes that appears as the head of lists and not the tail since order matters(will become clear below)

The linearisation of Third can be computed as follows:

    L(O)  := [O]  // the linearization(mro) of O(object), because O has no parents

    L(First)  :=  [First] + merge(L(O), [O])
               =  [First] + merge([O], [O])
               =  [First, O]

    // Similarly, 
    L(Second)  := [Second, O]

    L(Third)   := [Third] + merge(L(First), L(Second), [First, Second])
                = [Third] + merge([First, O], [Second, O], [First, Second])
// class First is a good candidate for the first merge step, because it only appears as the head of the first and last lists
// class O is not a good candidate for the next merge step, because it also appears in the tails of list 1 and 2, 
                = [Third, First] + merge([O], [Second, O], [Second])
// class Second is a good candidate for the second merge step, because it appears as the head of the list 2 and 3
                = [Third, First, Second] + merge([O], [O])            
                = [Third, First, Second, O]

Thus for a super() implementation in the following code:

class First(object):
  def __init__(self):
    super(First, self).__init__()
    print "first"

class Second(object):
  def __init__(self):
    super(Second, self).__init__()
    print "second"

class Third(First, Second):
  def __init__(self):
    super(Third, self).__init__()
    print "that's it"

it becomes obvious how this method will be resolved

Third.__init__() ---> First.__init__() ---> Second.__init__() ---> 
Object.__init__() ---> returns ---> Second.__init__() -
prints "second" - returns ---> First.__init__() -
prints "first" - returns ---> Third.__init__() - prints "that's it"

In python 3.5+ inheritance looks predictable and very nice for me. Please looks at this code:

class Base(object):
  def foo(self):
    print("    Base(): entering")
    print("    Base(): exiting")


class First(Base):
  def foo(self):
    print("   First(): entering Will call Second now")
    super().foo()
    print("   First(): exiting")


class Second(Base):
  def foo(self):
    print("  Second(): entering")
    super().foo()
    print("  Second(): exiting")


class Third(First, Second):
  def foo(self):
    print(" Third(): entering")
    super().foo()
    print(" Third(): exiting")


class Fourth(Third):
  def foo(self):
    print("Fourth(): entering")
    super().foo()
    print("Fourth(): exiting")

Fourth().foo()
print(Fourth.__mro__)

Outputs:

Fourth(): entering
 Third(): entering
   First(): entering Will call Second now
  Second(): entering
    Base(): entering
    Base(): exiting
  Second(): exiting
   First(): exiting
 Third(): exiting
Fourth(): exiting
(<class '__main__.Fourth'>, <class '__main__.Third'>, <class '__main__.First'>, <class '__main__.Second'>, <class '__main__.Base'>, <class 'object'>)

As you can see, it calls foo exactly ONE time for each inherited chain in the same order as it was inherited. You can get that order by calling .mro :

Fourth -> Third -> First -> Second -> Base -> object

Maybe there’s still something that can be added, a small example with Django rest_framework, and decorators. This provides an answer to the implicit question: “why would I want this anyway?”

As said: we’re with Django rest_framework, and we’re using generic views, and for each type of objects in our database we find ourselves with one view class providing GET and POST for lists of objects, and an other view class providing GET, PUT, and DELETE for individual objects.

Now the POST, PUT, and DELETE we want to decorate with Django’s login_required. Notice how this touches both classes, but not all methods in either class.

A solution could go through multiple inheritance.

from django.utils.decorators import method_decorator
from django.contrib.auth.decorators import login_required

class LoginToPost:
    @method_decorator(login_required)
    def post(self, arg, *args, **kwargs):
        super().post(arg, *args, **kwargs)

Likewise for the other methods.

In the inheritance list of my concrete classes, I would add my LoginToPost before ListCreateAPIView and LoginToPutOrDelete before RetrieveUpdateDestroyAPIView. My concrete classes’ get would stay undecorated.

Posting this answer for my future referance.

Python Multiple Inheritance should use a diamond model and the function signature shouldn’t change in the model.

    A
   / \
  B   C
   \ /
    D

The sample code snippet would be ;-

class A:
    def __init__(self, name=None):
        #  this is the head of the diamond, no need to call super() here
        self.name = name

class B(A):
    def __init__(self, param1='hello', **kwargs):
        super().__init__(**kwargs)
        self.param1 = param1

class C(A):
    def __init__(self, param2='bye', **kwargs):
        super().__init__(**kwargs)
        self.param2 = param2

class D(B, C):
    def __init__(self, works='fine', **kwargs):
        super().__init__(**kwargs)
        print(f"{works=}, {self.param1=}, {self.param2=}, {self.name=}")

d = D(name='Testing')

Here class A is object