Difference between filter and list comprehension

Question

I am using Python 3, my question is why is the output different?

print([x * x for x in range(2, 5, 2) if x % 4 == 0]) # returns [16]

q = [x * x for x in range(2, 5, 2)]
print(list(filter(lambda x: x % 4 == 0, q))) # returns [4, 16]

Answer 1

print(([x * x for x in range(2, 5, 2) if x % 4 == 0]))

here, range evaluates to [2,4] and only [4] is able to get past the if condition

q = ([x * x for x in range(2, 5, 2)])
print(list(filter(lambda x: x % 4 == 0, q)))

here, q contains x*x of each element, hence the list is [2*2, 4*4] = [4, 16], and both elements get past filter selector

Answer 2

Because q contains squares

In [2]: q
Out[2]: [4, 16]

and lambda x: x % 4 == 0 will return True for both of them:

In [3]: 4 % 4 == 0
Out[3]: True

In [4]: 16 % 4 == 0
Out[4]: True

The list comprehension squares numbers after performing the check, which fails for 2 (2 % 4 is 2):

In [5]: 2 % 4 == 0
Out[5]: False

Therefore, 2 * 2 = 4 won't be included in the list.

In short, if you want to get the same behaviour, modify your list comprehension to square numbers before computing the remainder:

[x * x for x in range(2, 5, 2) if pow(x, 2, 4) == 0] # [4, 16]
#                                   ↖ equivalent to (x ** 2) % 4