Chapter Eleven · Data That Knows What to Do

Objects and Classes

Every value in Python is an object: data bundled with the things it can do. This chapter shows how to define your own kind of object with a class, how dataclasses write the tedious parts for you, and, just as important, when a plain dictionary is the better choice.

Since Chapter 2 you have been using objects without the word: every time you wrote a dot, as in text.upper() or todo.append("milk"), you asked an object to do something. A string knows how to make an upper-case copy of itself; a list knows how to grow. This chapter lets you make your own kinds of object.

Three topics, deliberately few. Classes: how to define one, what self means, and how methods work. Dataclasses: the short way to write the most common kind of class, a record with named fields. And an honest rule of thumb for choosing between a class, a dataclass and the dictionaries you already know, because a beginner's most common mistake with classes is using them where nothing needed one.

Two ways to hold a record

A dictionary

Data only. Anything can be put in it, under any key.

A class

Data plus the things it can do, with a fixed shape.

Topics in This Chapter

Topic 01

Objects: Data That Knows What to Do

Every value is an object. Defining a class with __init__, what self refers to, methods that change an object, and a readable printout.

Hands-on

Topic 02

Dataclasses: The Short Form

A decorator that writes the constructor, the printout and equality for you from a list of fields. Defaults, methods, and frozen records.

Hands-on

Topic 03

When a Class Helps and When a Dict Is Enough

An honest rule of thumb: when a dictionary is enough, when a dataclass helps, when a class with methods earns its place, and how to convert between them.

Hands-on