Robust Python

Chapter 9. User-Defined Types: Data Classes

Data classes are user-defined types that let you group related data together. Many types, such as integers, strings, and enumerations, are scalar; they represent one and only one value. Other types, such as lists, sets, and dictionaries, represent homogeneous collections. However, you still need to be able to compose multiple fields of data into a single data type. Dictionaries and tuples are OK at this, but they suffer from a few issues. Readability is tricky, as it can be difficult knowing what a dictionary or tuple contains at runtime. This makes them hard to reason about when reading and reviewing code, which is a major blow to robustness.

When your data is hard to understand, readers will make incorrect assumptions and won’t be able to spot bugs as easily. Data classes are easier to read and understand, and the typechecker knows how to naturally handle them.

Data Classes in Action

Data classes represent a heterogeneous collection of variables, all rolled into a composite type. Composite types are made up of multiple values, and should always represent some sort of relationship or logical grouping. For example, a Fraction is an excellent example of a composite type. It contains two scalar values: a numerator and a denominator.

from fraction import Fraction
Fraction(numerator=3, denominator=5)

This Fraction represents the relationship between that numerator and denominator. The numerator and denominator are independent of each other; ...