Comprehensions and Iteration Tools

Introduction

A very common job is: take a collection, change or filter each item, and collect the results. With a loop it takes several lines. Python has a compact form called a comprehension that does it in one, and once you are used to it, it is easier to read.

From loop to comprehension

First the long way:

Now as a list comprehension. Read it as "n squared, for each n in numbers":

The pattern is [expression for item in collection].

Adding a filter

Add if at the end to keep only some items.

If you want to change values rather than drop items, use a conditional expression a if condition else b before the for:

Dictionary and set comprehensions

The same idea works with curly brackets.

Nested loops in a comprehension

Keep comprehensions short. If you need more than two for parts or a complicated condition, a normal loop is easier to read.

enumerate: index and value

Instead of counting yourself, let enumerate give you both.

zip: loop over two lists together

zip stops at the shortest list.

Sorting with a key

sorted accepts a key function that says what to sort by.

lambda p: p[1] is a tiny unnamed function that returns the second item of each pair.

any, all, sum, min and max

Try it yourself

Given a list of words, build a dictionary that maps each word longer than 3 letters to its length, using one comprehension. Then print the words sorted by length, longest first.

Summary

• [expr for x in items if cond] builds a list. Curly brackets make dictionary and set comprehensions.
• enumerate gives index and value, and zip loops over lists side by side.
• sorted(items, key=function) sorts by any rule. any and all test whole collections.
• Keep comprehensions short and readable. Use a normal loop for complicated logic.