Operator Overloading

Since operators call dunder methods, custom classes can define their own behavior for operators. This is called operator overloading.

Example: Custom Multiplication

Consider a scenario where multiplying a Horse by a Donkey should produce a Mule. This can be implemented by defining __mul__ on the Horse class:

class Mule:
    pass


class Donkey:
    pass


class CantBreedError(Exception):
    pass


class Horse:
    def __mul__(self, other: object):
        if isinstance(other, Donkey):
            return Mule()
        if isinstance(other, Horse):
            return Horse()
        raise CantBreedError("Horses can only breed with other horses or donkeys.")


# Usage
print(Horse() * Horse())   # <__main__.Horse object at 0x...>
print(Horse() * Donkey())  # <__main__.Mule object at 0x...>

When Python evaluates Horse() * Donkey(), it calls Horse().__mul__(Donkey()). The __mul__ method checks the type of the operand and returns the appropriate result.

Practical Applications

Operator overloading is used extensively in Python libraries:

• NumPy: Arrays support arithmetic operators for element-wise operations
• Pandas: DataFrames and Series support comparison operators for filtering
• pathlib: The / operator joins path components
• SQLAlchemy: Comparison operators build SQL WHERE clauses

By implementing dunder methods, custom classes can integrate seamlessly with Python's syntax, making code more readable and expressive.