Python data classes

from dataclasses import dataclass

@dataclass
class MyDataClass:
    name: str
    surname: str
    age: int

A data class is created using the @dataclass decorator. Type hinting is mandatory; Any can be used in place of a more specific data type.

The following hints are available for more complex datatypes: List[<type>], Tuple(<type>) and Dict(<type>, <type>).

Data class objects can be instantiated, printed and compared right out of the box. This just means that __init__(), __repr__() and __eq__() special methods are already implemented.

Providing arguments to the @dataclass decorator allows to customize which special method should or should not be implemented out of the box.

from dataclasses import make_dataclass

MyDataClass = make_dataclass('MyDataClass', ['name', 'surname', 'age'])

This is a named tuple like way of creating a data class.

@dataclass
class MyDataClass:
    name: str = 'John'
    surname: str = 'Doe'
    age: int

Simple default values for one or more fields can be provided at declaration time.

@dataclass
class MyDataClass:
    name: str = 'John'
    surname: str = 'Doe'
    age: int = field(default_factory = evaluate_age)

To provide more complex default values, use the field() specifier with a zero parameters callable as the argument for default_factory.

The field() specifier is used to customize each field of the data class individually.

Data classes support methods just like a regular class.

Special object methods can be overridden as well.

A data class can be declare as immutable by supplying the frozen = True argument to the @dataclass decorator.

Note that if the data class contains mutable fields (ie: lists) they might still change.

@dataclass
class MyDataClass:
    ...

@dataclass
class MySubDataClass(MyDataClass):
    ...

Data classes can be sub-classed as usual.

Note: in case the base class employs fields with a default value, provide default values for all fields of the subclass as well.