Python type() Function with Three Arguments

The type() function is commonly used with a single argument – the variable whose data type we want to check. It returns an existing type object – the type of the input variable. For instance, if passed an int like 123, it returns the int type.

However, there is another variant with three arguments:

type(name, bases, dict)

It creates a new type object. This can be used to create new classes dynamically.

Inputs

The three inputs are:

• name (str) = name of the new class (it becomes its __name__ attribute).
• bases (tuple) = base classes of the new class (it becomes the __bases__ attribute); if empty the base class is object.
• dict (dict) = attribute and method definitions (the __dict__ attribute).

Example

For example, we can use type() to create a new class named Number, subclass of int:

Number = type(
    'Number', 
    (int,), 
    {'is_even': lambda self: self % 2 == 0}
)

• The first argument is the name of the new class, which is 'Number' (the input is a string – with quotes).
• The second argument is a tuple with a single item – the int class, which is the base class of the new Number class. The items are classes/types, not strings, so no quotes.
• The third argument is a dict with a single item – the definition of a new method named 'is_even', which checks whether the number is even. Dict keys are attribute or method names as strings – with quotes.

Using the new class

Now we can create an instance of the new class Number and use its method is_even to check if a number is even:

>>> x = Number(6)
>>> x.is_even()
True
>>> y = Number(7)
>>> y.is_even()
False

If we check the data type using type() function (with a single argument), we get the new Number type.

>>> type(y)

We can also use isinstance() to check that int is indeed a base class of Number.

>>> isinstance(y, int)
True

Official documentation

https://docs.python.org/3/library/functions.html#type