Note: Python 3 is used as the operating environment here, and the example is taken from Chapter 8 of "python cookbook".
In python, if a subclass wants to initialize the parent class, there are two main methods. The first is to directly pass the parent class name, and the second is to use the super method. There is no difference between the two in single inheritance, but there are some subtle differences that need to be paid attention to in multiple inheritance. Example explanation is the last word!
1. The case of using the parent class name:
Python code
class Base:
def __init__(self):
print('Base.__init__')
class A (Base):
def __init__(self):
use using using using use using ‐ ‐ use using out out out out out of ’s to def __init__(self)’s to be __init__(self’ be-to-be--
Base.__init__(self) " B.__init__(self)
print( 'C.__init__')
At this time, instantiating class C will output the following:Python code
>>> c = C()
Base.__init__
A.__init__
Base .__init__
B.__init__
C.__init__
>>>
It can be seen that the Base class is called twice. This must not be the result we want in many cases, so we can consider using the super method at this time.
Python code
class Base:
def __init__(self):
print('Base.__init__')
class A (Base):
def __init__(self):
super().__init__( )
) to super() here
At this time, the output of Class C is instantiated at this time:
python code & & gt; & gt; & gt; c = c () base .____ B. __init__ A.__init__ C.__init__ >>>It can be seen that the Base class is only called once! But unfortunately, this is not what prompted me to write this blog record, because if you observe carefully, although the Base class is indeed called only once as expected, have you noticed that "B.__init__" is output first and then Just output "A.__init__"? And why does this cause Base to be initialized only once? You must be a little confused too, right? In fact, all this is "blamed" on the calling process of super during multiple inheritance. When python implements a class (not just inheritance), it will generate a method to generate a parsing order list, which can be viewed through the class attribute __mro__, as in this example:
Python code >> ;> C.__mro__ (
So when searching for a property or method, it will iterate through each class according to this list until it finds the first class that matches this property or method. When using super in inheritance, the interpreter will search for the next class on the list every time it encounters super until it no longer encounters super or the list is traversed, and then returns layer by layer similar to recursion. Therefore, the search process in this example is: encounter super in C --> search for the next class in the list, that is, A --> encounter super again in A, search for B --> super reappears in B, search Base --> Initialize the Base class and return recursively.
In order to better explain the process, please comment out the super line of class B now:
Python code
class B(Base):
def __init__(self):
#super().__init__( )
) to super() here
'Print (' C.__ init__ ')Class C again, output as follows:
pythonn code & & gt; & gt; c = c () b .__
C.__init__
Base class no longer produces output! Why? Because there is no super in B, the list is blocked from searching for the Base class, so Base is not initialized!