Python

t = {
    'a': '1',
    'b': '2',
    'c': '3',
}


class Counter(dict):

    def __missing__(self, key):
        return None
c = Counter(t)
print(c['d'])

There will be:

None

Change return value:

Python

t = {
    'a': '1',
    'b': '2',
    'c': '3',
}


class Counter(dict):

    def __missing__(self, key):
        return key
c = Counter(t)
print(c['d'])
print(c)

There will be:

d
{'c': '3', 'a': '1', 'b': '2'}

The fifth solution
Using the collections.defaultdict([default_factory[...]]) object, which is actually inherited from the dict and is actually used with the missing__ __() method, its default_factory argument is passed to the missing__() method,
if default_factory is None, it will be the same as dict, it will trigger KeyError error, as follows:

Python

import collections
t = {
    'a': '1',
    'b': '2',
    'c': '3',
}
t = collections.defaultdict(None, t)
print(t['d'])

There will be:

KeyError: 'd'

But if you really want to return None is not out of the question:

Python

import collections
t = {
    'a': '1',
    'b': '2',
    'c': '3',
}

def handle():
    return None
t = collections.defaultdict(handle, t)
print(t['d'])

There will be:

None

If the default_factory parameter is a data type, it will return its default value, such as:

Python

import collections
t = {
    'a': '1',
    'b': '2',
    'c': '3',
}
t = collections.defaultdict(int, t)
print(t['d'])

There will be:

0

Such as:

Python

import collections
t = {
    'a': '1',
    'b': '2',
    'c': '3',
}
t = collections.defaultdict(list, t)
print(t['d'])

There will be:

[]

Note:
if dict contains dict, key nested to get value, if a middle key0 does not exist, then all the above methods are invalid, and 1 KeyError2 will be triggered:

Python

import collections
t = {
    'a': '1',
    'b': '2',
    'c': '3',
}
t = collections.defaultdict(dict, t)
print(t['d']['y'])

There will be:

KeyError: 'y'