using **kwargs in __init__

Question

I am running into the following problem:

class Student():
    def __init__(self,firstname, lastname, **kwargs):
        self.firstname = firstname
        self.lastname = lastname

student1 = Student("John" ,"Johnson" , {"Nationality": "Middle-Earth" , "Sports": ["volleyball","skiing"]})

I want to be able to create students that all have a first and a last name, but only some have information about nationality or sports. (Actually I am experimenting with reconstructing objects from json files.) So I want to turn kwarg key into instance variable names, like this pseudo-code

for key,value in kwargs:
   self.key = value

the above is pseude-code, but is there a way to do this in correct python? in other words: derive instance variable names from kwargs rather than all predefined.

I know this won't result in errorproof code etc..., but I want to know if it can be done.

Answer 1

You can use this in the __init__ method:

for k, v in kwargs.items():
    setattr(self, k, v)

However, this is not recommended for code than needs to be reusable. If there are typos in the keyword parameters, you won't catch them. Code-checking tools such as pylint will not recognize class attributes initialized like that. You can use the following pattern (especially if there are many kewyord arguments):

class Student:
    def __init__(self, firstname, lastname, **kwargs):
        self.firstname = str(firstname)
        self.lastname = str(lastname)
        self.nationality = None
        self.sports = None

        for k, v in kwargs.items():
            if k in self.__dict__:
                setattr(self, k, v)
            else:
                raise KeyError(k)

# this works
s=Student('John', 'Doe', sports=['socker'])

# this will give an error
s=Student('John', 'Doe', sprots=['socker'])