Python Comprehensions: A step by step Introduction

List Comprehensions are a special kind of syntax that let us create lists out of other lists (Wikipedia, The Python Tutorial). They are incredibly useful when dealing with numbers and with one or two levels of nested for loops, but beyond that, they can become a little too hard to read.

In this article, we are going to make some For Loops and rewrite them, step by step, into Comprehensions.

Basics

The truth is, List Comprehensions are not too complex, but they are still a bit difficult to understand at first because they look a little weird. Why? Well, the order in which they are written is the opposite of what we usually see in a For Loop.

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> for n in names:
...     print(n)
# Charles
# Susan
# Patrick
# George
# Carol

To do the same with a List Comprehension, we start at the very end of the loop:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> [print(n) for n in names]
# Charles
# Susan
# Patrick
# George
# Carol

Notice how we inverted the order:

• First, we have what the output of the loop will be [print(n) ...].
• Then we define the variable that will store each of the items and point at the List, Set or Dictionary we will work on [... for n in names].

Creating a new List with a Comprehension

This is the primary use of a List Comprehension. Other usages may result in a hard-to-read code for you and others.

This is how we create a new list from an existing collection with a For Loop:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']

>>> new_list = []
>>> for n in names:
...     new_list.append(n)

>>> print(new_list)
# ['Charles', 'Susan', 'Patrick', 'George', 'Carol']

And this is how we do the same with a List Comprehension:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> new_list = [n for n in names]
>>> print(new_list)
# ['Charles', 'Susan', 'Patrick', 'George', 'Carol']

The reason we can do this is that a List Comprehension standard behavior is to return a list:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> [n for n in names]
# ['Charles', 'Susan', 'Patrick', 'George', 'Carol']

Adding Conditionals

What if we want new_list to have only the names that start with C? With a For Loop, we would do it like this:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']

>>> new_list = []
>>> for n in names:
...     if n.startswith('C'):
...         new_list.append(n)

>>> print(new_list)
# ['Charles', 'Carol']

In a List Comprehension, we add the if statement at its end:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> new_list = [n for n in names if n.startswith('C')]
>>> print(new_list)
# ['Charles', 'Carol']

A lot more readable.

Formatting long List Comprehensions

This time, we want new_list to have not only the names that start with a C but also those that end with an e and contain a k:

>>> names = ['Charles', 'Susan', 'Patrick', 'George', 'Carol']
>>> new_list = [n for n in names if n.startswith('C') or n.endswith('e') or 'k' in n]
>>> print(new_list)
# ['Charles', 'Patrick', 'George', 'Carol']

That is quite messy. Fortunately, it is possible to break Comprehensions in different lines:

new_list = [
    n
    for n in names
    if n.startswith("C")
    or n.endswith("e")
    or "k" in n
]

Set and Dict Comprehensions

If you learned the basics of List Comprehensions… Congratulations! You just did it with Sets and Dictionaries.

Set comprehension

>>> my_set = {"abc", "def"}

>>> # Here, we create a new set with uppercase elements with a for loop
>>> new_set = set()
>>> for s in my_set:
...    new_set.add(s.upper())
>>> print(new_set)
# {'DEF', 'ABC'}

>>> # The same, but with a set comprehension
>>> new_set = {s.upper() for s in my_set}
>>> print(new_set)
# {'DEF', 'ABC'}

Dict comprehension

>>> my_dict = {'name': 'Christine', 'age': 98}

>>> # A new dictionary out of an existing one with a for loop
>>> new_dict = {}
>>> for key, value in my_dict.items():
...     new_dict[key] = value
>>> print(new_dict)
# {'name': 'Christine', 'age': 98}

# Using a dict comprehension
>>> new_dict = {key: value for key, value in my_dict.items()}  # Notice the ":"
>>> print(new_dict)
# {'name': 'Christine', 'age': 98}

Recommended Article: Python Sets: What, Why and How .

Conclusion

Every time I learn something new, there is this urge to use it right away. When that happens, I force myself to stop and think for a moment… Should I change this big, nested and already messy looking For Loop to a List Comprehension? Probably not.

Readability counts. The Zen of Python.