Is it necessary to increase the complexity?

Question

Python

Please tell me, I understand correctly that in order to determine the complexity of the following algorithm, you need O (n) * O (n) = O (n ^ 2)

#!/usr/bin/env python3

def reverse_letter(string):
    return ''.join([l for l in reversed(string) if l.isalpha()])

I do this because reverse will iterate over all elements in the worst case, and for will also iterate over all elements. Thus, the final complexity will be quadratic

Answer 1

No. Now, if you had two nested loops on string, then yes. And here the complexity "adds up", because first O (n) - reverse, then O (n) - creating a cycle, then O (n) - join. But since no one needs constants in complexity - it's only O (n)
But, of course, you won't believe it, so here's the proof for you. Time grows linearly with N

Answer 2

If I understand correctly, reversed will only happen once. In pseudocode, you can write it like this:

newStr = reversed(string);
for sym in newStr; // ..

That is, a linear pass is used here twice, and not a linear one inside a linear one, which means that there is no need to multiply - the complexity is linear.