More Python Comprehensions

Table of Contents

Python comprehension == one_liner.

Basic L1
#

values = []
for x in range(1,11):
    values.append(x)

The block of code above can be simplified as:
values = [ x for x in range(1,11) ]

evens = []
for n in range(1,11):
    is_even = n % 2 == 0
    if is_even:
        evens.append(number)

The block of code above can be simplified as:
evens = [ n for n in range(1,11) if n % 2 == 0 ]

Level 2
#

names = [ 'jane', 'jenny', 'jim', 'jimmy', 'jimny', 'jone', 'june' ]
valid = []

for name in names:
    if len(name) <= 1:
        continue

    if name[0] != 'j':
        continue

    if name[-1] != 'j':
        continue

    valid.append(name)

The block of code above can be simplified as:
valid = [ name for name in names if len(name) >= 2 if name[0] =='j' if name[-1] == 'y' ]

matrix = [['a','b','c'],['d','e','f'],['g','h','i']]
flat = []

for y in matrix:
    for x in y:
        flat.append(x)

The block of code above can be simplified as:
flat = [ x for y in matrix for x in y ]

types = []

for n in range(2,12):
    if n % 2 ==0:
        types.append('E')
    else:
        types.append('O')

The block of code above can be simplified as:
types = [ "E" if n % 2 == 0 else "O" for n in range(2,12) ]

3D-Matrix
#

>>> mm = []
>>>
>>> for a in range(5):
...     m1 = []
...     for b in range(4):
...         m2 = []
...         for c in range(3):
...             m2.append(c)
...         m1.append(m2)
...     mm.append(m1)
...
>>> mm
[
    [[0, 1, 2], [0, 1, 2], [0, 1, 2], [0, 1, 2]],
    [[0, 1, 2], [0, 1, 2], [0, 1, 2], [0, 1, 2]],
    [[0, 1, 2], [0, 1, 2], [0, 1, 2], [0, 1, 2]],
    [[0, 1, 2], [0, 1, 2], [0, 1, 2], [0, 1, 2]],
    [[0, 1, 2], [0, 1, 2], [0, 1, 2], [0, 1, 2]]
]

The block of code above can be simplified as:
mmm = [[[x for x in range(3)] for _ in range(4)] for _ in range(5) ]

Function
#

def square(x):
    return x**2

square_num = []
for x in range(1,11):
    square_num.append(square(x))

The block of code above can be simplified as:
square_num = [ square(x) for x in range(1,11) ]

Dictionary
#

pairs = [ ('a', 1), ('b', 2), ('c', 3) ]

my_dict = { k:v*2 for k,v in pairs }

Set
#

Python set() uses the same {} as dict() as long as there is no key.

Below here is a typical example where many will use it to remove duplicate values in a list().

nums = [1,1,2,2,2,3,4,4,5,6,6]

uniqsqs = { x**2 for x in nums }

Generator
#

sum_of_sqs = sum(x**2 for x in range(1_000_000))

The above code block will generate a sum of all the squares from the number 0 to 1 millioni, and just give me the end result (without storing all the different values).

This means it is different than sum([x**2 for x in range(1_000_000)]). This will generate a list of 1 million values before sum them up.

Using a generator, it is more efficient in term of memory.

Summary
#

Thanks to Tim with the excellent tutorial on 10 Python Comprehensions.

I’ve modified some codes to suit my own study. Below are the quick summary.

# Basic L1
values = [ x for x in range(1,11) ]

evens = [ n for n in range(1,11) if n % 2 == 0 ]

# Level L2
names = [ 'jane', 'jenny', 'jim', 'jimmy', 'jimny', 'jone', 'june' ]
valid = [ name for name in names if len(name) >= 2 if name[0] =='j' if name[-1] == 'y' ]

matrix = [['a','b','c'],['d','e','f'],['g','h','i']]
flat = [ x for y in matrix for x in y ]

types = [ "E" if n % 2 == 0 else "O" for n in range(2,12) ]

# 3D-Matrix
mmm = [[[x for x in range(3)] for _ in range(4)] for _ in range(5) ]

# Function
square_num = [ square(x) for x in range(1,11) ]

# Dictionary
pairs = [ ('a', 1), ('b', 2), ('c', 3) ]
my_dict = { k:v*2 for k,v in pairs }

# Set
nums = [1,1,2,2,2,3,4,4,5,6,6]
uniqsqs = { x**2 for x in nums }

# Generator
sum_of_sqs = sum(x**2 for x in range(1_000_000))