Correct understanding and usage analysis of super function in Python programming

When a subclass needs to call a method of the parent class, before python2.2, the method of the class is directly called using the class name, that is, a non-bound class method, and the own object self is passed in as a parameter.

class A(object): 
  def say(self): 
    print 'I am A' 
 
class B(A): 
  def say(self): 
    print 'I am B' 
    A.say(self) 
 
b = B() 
b.say()

Output

I am B
I am A

This works great , but there is a problem. When the parent class changes its name, these explicit calls to the parent class must be corrected one by one. The coupling between the subclass and the parent class is relatively high.
So the super() function was introduced after Python 2.2 to avoid hard coding and no need to care about the name of the parent class.
Using the super() function, the above code can be written as follows.

class B(A): 
  def say(self): 
    print 'I am B' 
    super(B,self).say()

After python3.0, improvements have been made. The super() function does not need to pass parameters, that is, the above line of code directly super(). Just say() will do.

Things to note:

• super can only be used in new-style classes.

• Super has a problem with multiple inheritance. If a subclass inherits multiple parent classes, then super calls the method of the first parent class.

• Don’t mix these two methods of calling parent class methods, either use unbound class methods or both use super. Otherwise, it may not be called or may be called multiple times.

BUT:
Don’t think of the parent class when you think of super! super refers to the next class in MRO!
When talking about super, I think of the parent class. This is a mistake that beginners can easily make, and it was also a mistake I made back then.

def super(cls, inst):
  mro = inst.__class__.mro()
  return mro[mro.index(cls) + 1]

The two parameters cls and inst do two things respectively:
1. inst is responsible for generating the MRO list
2. Locate the index in the current MRO through cls, and return mro[index + 1]
These two things are the essence of super, you must remember it!
MRO stands for Method Resolution Order, which represents the order of class inheritance.

For example:

class Root(object):
  def __init__(self):
    print("this is Root")

class B(Root):
  def __init__(self):
    print("enter B")
    # print(self) # this will printsuper(B, self).__init__()
    print("leave B")

class C(Root):
  def __init__(self):
    print("enter C")
    super(C, self).__init__()
    print("leave C")

class D(B, C):
  pass

d = D()
print(d.__class__.__mro__)

Output

enter B
enter C
this is Root
leave C
leave B
(,,,,)

After knowing that super has no real relationship with the parent class, it is not difficult for us to understand why the next sentence of enter B is enter C instead of this is Root (if you think super represents "calling the method of the parent class", you will I take it for granted that the next sentence should be this is Root). The process is as follows, in the __init__ function of B:

super(B, self).__init__()

First, we get self.__class__.__mro__. Note that self here is the instance of D. Instead of B's ​​

(

, then, use B to locate the index in the MRO and find the next one. Obviously B's next one is C. So, we call C's __init__ and type enter C.

By the way, why B's __init__ will be called: Because D does not define __init__, so we will find the next class in MRO to see if it has __init__ defined, that is, to call B's __init__.

In fact, the logic of all this is still very clear. The key is to understand what super does.

The above is the correct understanding and usage analysis of the super function in Python programming. For more related content, please pay attention to the PHP Chinese website (www.php.cn)!