What is the most pythonic way to sort a dict with list values?

Question

I am creating a dictionary with values as a list of integers. What is the pythonic way to sort the values for each key? In my approach I am overwriting each key value after sorting.

Is there a way to maintain the sorting on insertion into the dictionary or is it better to store first and then sort the values efficiency wise?

my_dict = {
'first': [3,6,99],
'second': [1,-10,100],
'third': [5,0,23,67,29]
}

# my approach
for key, values in my_dict.items():
    my_dict[key] = sorted(values)

Ouput:

my_dict = {
'first': [3,6,99],
'second': [-10,1,100],
'third': [0,5,23,29,67]
}

Answer 1

You were almost there, you want to sort the lists in place, so uselist.sort :

my_dict = {
    'first': [3,6,99],
    'second': [1,-10,100],
    'third': [5,0,23,67,29]
}

for val in my_dict.values():
    val.sort()

Live demo

Check this sorting how to guide

Answer 2

Solution:

pythonic code makes use of programming patterns such as list comprehensions or generator expressions, attempt to crowbar in patterns more commonly used in C or java. Loops are particularly common examples of this.Pythonic code also makes use of inline sintaxes as mentioned by PEP20 ( Beautiful is better than ugly. , Flat is better than nested. )

Firstly, your code is good but to be more pythonic you should make use of sytaxes such as list comprehnsions/dict comprehensions. I will cover the most efficient way that has the least lines of code using a dict comprehension instaed of a for loop to be more pythonic even hough for loops are considered to be pyhonic list & dict comprehensions are more pythonic(they are flat syntaxes).

Here is the code:

my_dict = {
'first': [3,6,99],
'second': [1,-10,100],
'third': [5,0,23,67,29]
}

sorted_dict = {key:sorted(value) for key, value in my_dict.items()}
print(sorted_dict)

Output:

{'first': [3, 6, 99], 'second': [-10, 1, 100], 'third': [0, 5, 23, 29, 67]}

Answer 3

Local sorting does not need to assign a value to the key again.

import random
import time
from copy import deepcopy


def time_it(func):
    def inner():
        start = time.time()
        func()
        end = time.time()
        print('time used:{} second'.format(end - start))

    return inner


def random_l(n):
    return [random.randint(0, 100) for _ in range(n)]


my_dict = {
    'first': random_l(5),
    'second': random_l(5),
    'third': random_l(5)
}
my_dict1 = deepcopy(my_dict)
my_dict2 = deepcopy(my_dict)


@time_it
def foo():
    print(my_dict1)
    for key, values in my_dict1.items():
        my_dict1[key] = sorted(values)
    print(my_dict1)


@time_it
def bar():
    print(my_dict2)
    for key, values in my_dict2.items():
        sorted(my_dict2[key])
    print(my_dict2)


if __name__ == '__main__':
    foo()
    bar()

{'first': [28, 96, 52, 57, 93], 'second': [17, 89, 76, 30, 36], 'third': [73, 32, 9, 90, 81]}
{'first': [28, 52, 57, 93, 96], 'second': [17, 30, 36, 76, 89], 'third': [9, 32, 73, 81, 90]}
time used:5.3882598876953125e-05 second
{'first': [28, 96, 52, 57, 93], 'second': [17, 89, 76, 30, 36], 'third': [73, 32, 9, 90, 81]}
{'first': [28, 96, 52, 57, 93], 'second': [17, 89, 76, 30, 36], 'third': [73, 32, 9, 90, 81]}
time used:1.621246337890625e-05 second

If you want a feature like sortedList sortedcontainers is convenient, or you can use the insort function of the bitsect module