How to convert this my code into a list comprehension

Question

I've written this code so that it generates 4 random ints ranging from 1-6 and then remove the smallest number and add it to a list that is returned.

I was reading around and found that list comprehensions are the more "pythonic" solution instead of these small for range loops. I would like to know how to write this code as a list comprehension and any help would be greatly appreciated.

stats = []

for stat in range(6):
    score = [random.randint(1, 6) for n in range(4)]
    score.remove(min(score))
    stats.append(sum(score))

return stats

Answer 1

In Python 3.7 or less, you can do this using the following combination of list comprehensions + generator expression:

stats = [sum(score) - min(score) for score in ([random.randint(1, 6) for n in range(4)] for stat in range(6))]

In Python 3.8 (still in beta), you can do it in a simpler way thanks to the new walrus assignment operator:

stats = [sum(score := [random.randint(1, 6) for n in range(4)]) - min(score) for stat in range(6)]

You can try it here.

Testing both approaches:

Comprehensions (<=Python 3.7):

import random

random.seed(1)

stats = [sum(score) - min(score) for score in ([random.randint(1, 6) for n in range(4)] for stat in range(6))]

print(stats)

Output:

[10, 12, 12, 12, 15, 14]

Comprehensions + walrus (Python 3.8):

import random

random.seed(1)

stats = [sum(score := [random.randint(1, 6) for n in range(4)]) - min(score) for stat in range(6)]

print(stats)

Output:

[10, 12, 12, 12, 15, 14]

Answer 2

Here is my attempt:

I define a function to generate score before using the list comprehension.

import random
stats = []

def make_rand_score():
    score = [random.randint(1, 6) for n in range(4)]
    score.remove(min(score))
    return score
stats = [make_rand_score() for i in range(5)]

print(stats)

Answer 3

Try with this:

from random import randint
[min([randint(1, 6) for _ in range(4)]) for _ in range(6)]

However this gets a little complex and hard to read so maybe it's not the best solution.