Question

I’m trying to solve a codechef beginner problem – Enormous Input Test. My code

a,b = [ int(i) for i in raw_input().split()]
print [input()%b==0 for i in range(a)].count(True)

gets timed out. Another solution, which uses basic for-loops, seems to be working fine.

I believe that list comprehension is quicker than basic for – loops. Then why is the former slower? Also will using generators in this case reduce the memory used and perform the computation faster, if so how can I do it?

Answer 1

Why do you believe that list comprehension is quicker than basic for loops? (Hint: they are both implemented using the same underlying instructions.)

Your code will be executed in some manner like this:

a, b = ...
temp = []
for i in range(a):
    temp.append(int(raw_input()) % b == 0)
print temp.count(True)

As you can see, it creates a large list in memory, iterates over it to create a second list, and then iterates over the second list to create a count. The list does not ever need to be created.

a, b = ...
count = 0
for i in xrange(a):
    if int(raw_input()) % b == 0:
        count += 1
print count

Some compilers are capable of optimizing hylomorphisms to remove the intermideate list, but I know of no Python implementation capable of this. So you are stuck optimizing by hand.

Note: Do not use input in Python 2.x, unless you know what you are doing. I have changed the code to use int(raw_input()) because that is safe, whereas input() is dangerous.