When to use a list comprehension vs a for loop

Question

Is there a general rule for when a list comprehension is preferred over a for loop? For example:

A single for loop is great when converted to a comprehension:

l = []
for word in sentence:
    l.append(word)

# [word for word in sentence]

A double one might be too:

l = []
for word in sentence:
    for letter in word:
        l.append(letter)

# [letter for word in sentence for letter in word]

However, I think the readability gets pretty poor after that. For example, with ifs:

l = []
for word in sentence:
    if word.startswith('u'):
        for letter in word:
            if letter in ('a', 'b', 'c', 'o'):
                l.append(letter)

# [letter for word in sentence if word.startswith('u') for letter in word if letter in ('a', 'b', 'c', 'o')]

Is there a suggestion for the complexity of a for loop pattern where it would be worse trying to put it into a list comprehension?

I would write it both ways, and test them against each other to verify correctness. Then I would put it aside for a few days. Whichever one is more readable, without being fresh in my mind, would stay. — Tom Zych
– Tom Zych, Commented Mar 11, 2020 at 0:36
both are great ways, but it all comes down to preference and who you are presenting the code to. the for loop approach does shows a more 'methodical' approach as in, everything is clear as is, the comprehension way be it for list or dicts is great for getting the job done quickly. — Data_Is_Everything
– Data_Is_Everything, Commented Mar 11, 2020 at 0:43
On the topic of readability, many people dislike using lower case l for list because it looks too much like 1 :) — hilberts_drinking_problem
– hilberts_drinking_problem, Commented Mar 11, 2020 at 1:01

Vasili Syrakis · Accepted Answer · 2020-03-11 00:47:17Z

7

The answer depends on your opinion. However, since I do remember a specific piece of advice of this from an author of a book that is well regarded in the Python community, I can share the following excerpt from the book 'Fluent Python' by Luciano Ramalho:

A for loop may be used to do lots of different things: scanning a sequence to count or pick items, computing aggregates (sums, averages), or any number of other processing tasks. [...] In contrast, a listcomp is meant to do one thing only: to build a new list.

Of course, it is possible to abuse list comprehensions to write truly incomprehensible code. I’ve seen Python code with listcomps used just to repeat a block of code for its side effects.

If you are not doing something with the produced list, you should not use that syntax.

Also, try to keep it short. If the list comprehension spans more than two lines, it is probably best to break it apart or rewrite as a plain old for loop. Use your best judgment: for Python as for English, there are no hard-and-fast rules for clear writing.

answered Mar 11, 2020 at 0:47

Vasili Syrakis

9,6912 gold badges43 silver badges59 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Alistair Over a year ago

I'd agree with this, only thing to add that I haven't seen mentioned yet is: for those cases where nested list comprehension is what you need, you can define each step as a function to improve readability

Ukrainian-serge · Accepted Answer · 2020-03-11 00:39:27Z

2

According to this, performance can be gained.

import timeit

def squares(size):
    result = []
    for number in range(size):
        result.append(number*number)
    return result

def squares_comprehension(size):
    return [number*number for number in range(size)]

print(timeit.timeit("squares(50)", "from __main__ import squares", number = 1_000_000))
print(timeit.timeit("squares_comprehension(50)", "from __main__ import squares_comprehension", number = 1_000_000))

5.4292075
4.1652729000000015

Hope this helps.

answered Mar 11, 2020 at 0:39

Ukrainian-serge

8547 silver badges12 bronze badges

1 Comment

Vasili Syrakis Over a year ago

Yep, list comprehensions have some optimizations which makes them faster than for loops, and also map+filter.

hilberts_drinking_problem · Accepted Answer · 2020-03-11 00:58:17Z

1

I'd add that you could always work with generators and iterators. There is no major memory overhead, but the performance may suffer unless the innermost operation is relatively costly.

from itertools import chain
sentence = "flat is better than nested"

words = sentence.split()
f_words = (w for w in words if w.startswith('f'))
f_chars = chain(*f_words)
good_chars = [c for c in f_chars if c in 'abco']

answered Mar 11, 2020 at 0:58

hilberts_drinking_problem

11.6k3 gold badges25 silver badges55 bronze badges

Collectives™ on Stack Overflow

When to use a list comprehension vs a for loop

3 Answers 3

1 Comment

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related