CODESARRAY

Lesson 9, by Codesarray, , 4 min read

Python list comprehensions: lists, dicts, sets, if, any and all

Build a Python list, dictionary or set from another one in a single line, filter it with if, test it with any and all, and know when a plain loop is clearer.

A Python list comprehension builds a new list from an old one in a single line. By the end of this page you will write list, dictionary and set comprehensions, filter them with if, test a whole collection with any and all, and know when a plain loop is the clearer choice.

List comprehensions

Inside square brackets you write the expression first, then for, a name, in, and the source. [n * n for n in nums] gives the square of every number in nums. It does the same job as a loop that starts with an empty list and calls append on every pass, but it fits on one line and reads left to right.

Filtering with if

Add an if at the end to keep only some items. [n for n in nums if n % 2 == 0] keeps the even numbers. Each item is tested first, and only the ones that pass are collected.

Dictionary and set comprehensions

Curly brackets build other types. With a colon, key then value, you get a dictionary comprehension: {w: len(w) for w in words} maps every word to its length. Without a colon you get a set comprehension, and duplicates vanish. The letters of "hello" give four unique letters, not five.

any and all

any is True if at least one item passes the test. all is True only if every item passes. Write the test inside the brackets of the call, with no square brackets of its own: any(n > 4 for n in nums). Python checks the items one at a time and stops as soon as the answer is known.

Comprehension or loop?

A comprehension is for building one collection from another. If you need several steps, a side effect such as print, or a nested condition that takes a moment to read, use a plain loop. Short and clear beats clever.

nums = [1, 2, 3, 4, 5]

squares = [n * n for n in nums]
print(squares)

evens = [n for n in nums if n % 2 == 0]
print(evens)

lengths = {w: len(w) for w in ["red", "green"]}
print(lengths)

letters = {c for c in "hello"}
print(len(letters))

print(any(n > 4 for n in nums))
print(all(n > 4 for n in nums))
[1, 4, 9, 16, 25]
[2, 4]
{'red': 3, 'green': 5}
4
True
False

Line 1 is the list of numbers. Line 3 squares each one with a list comprehension and line 4 prints the result. Line 6 keeps only the evens, and line 7 prints them. Line 9 builds a dictionary from words to their lengths, and line 10 prints it. Line 12 builds a set of letters, and line 13 prints how many there are: four, because the two l characters collapse into one. Line 15 asks whether any number is bigger than 4, which is True because of 5. Line 16 asks whether all of them are, which is False because 1 is not.

What you learned

  • A list comprehension builds a list in one line: [expression for name in source].
  • An if at the end filters which items are kept.
  • Curly brackets with a colon make a dictionary, and without a colon make a set.
  • any and all test a whole collection and stop as soon as the answer is known.
  • When it gets complicated, a plain loop is the better choice.

Common questions

Is a list comprehension faster than a for loop? Usually a little, because Python does not look up and call append on every pass. Choose it for readability first; the speed difference rarely matters.

Can I use if and else in a comprehension? Yes, but the order changes. A filter goes at the end: [n for n in nums if n > 2]. A choice between two values goes at the front: [n if n > 2 else 0 for n in nums].

Can I write one comprehension inside another? Yes. [x for row in grid for x in row] flattens a list of lists. If it takes more than a moment to read, write the loops out instead.

How do I make an empty set? {} is an empty dictionary. Use set() for an empty set.

Does the loop variable leak out of the comprehension? No. After [k for k in nums], using k raises a NameError, because the name lives only inside the comprehension.

This lesson follows on from Python functions. Next in the series: errors and exceptions.

This article goes with the video Comprehensions. Watch it on YouTube.