Is it alright to call len() in a loop's conditional statement?

Question

In C, it is considered bad practice to call strlen like this:

for ( i = 0; strlen ( str ) != foo; i++ )
{
    // stuff
}

The reason, of course, is that it is inefficient since it "counts" the characters in a string multiple times.

However, in Python, I see code like this quite often:

for i in range ( 0, len ( list ) ):
    # stuff

Is this bad practice? Should I store the result of len() in a variable and use that?

Answer 1

In Python, a for loop iterates through a list-like object, it doesn't have a conditional statement that is checked each time. To illustrate, the following two loops are functionally equivalent; the while loop is a direct translation of for (i=0; i< n; i++) { ... }, while the for loop is the Pythonic way of doing it:

i = 0
while i < n:
   # do some actions using i
   i += 1

for i in range(n):
   # do some actions using i

In the for loop, range(n) is evaluated before the loop starts, and then its return value is used in the loop. In other words, range(n) is evaluated once (and only once) to get the list* of values that i should take.

So in the specific example given, Python sees for i in range(0, len(list)), evaluates range(0, len(list)) (so len is called once) and stores the output (it doesn't call range (or len) again).

(* in Python 3.x range doesn't actually return a list, but when using it in a for loop there is no difference in functionality)

Answer 2

Writing for i in range(len(whatever)) is poor style in Python. As others have mentioned, direct iteration is available via for element in whatever. However, people keep doing it anyway for two reasons:

1. They're C programmers and haven't wrapped their heads around direct iteration.

2. They want to modify the list while iterating through it, or otherwise need the indices. If this is the case, the Pythonic way to do it is the enumerate function: for i, element in enumerate(whatever).

Answer 3

This is fine in python because the range function creates a list that the for loop subsequently uses. Therefore len is only called once. Also, you don't need the extra 0 there because range starts counting from 0 if no extra options are there.

As a side note, you'll usually want to use xrange instead since it creates a generator that is evaluated lazily.