Question

So, in Python (though I think it can be applied to many languages), I find myself with something like this quite often:

the_input = raw_input("what to print?\n")
while the_input != "quit":
    print the_input
    the_input = raw_input("what to print?\n")

Maybe I’m being too picky, but I don’t like how the line the_input = raw_input("what to print?\n") has to get repeated. It decreases maintainability and organization. But I don’t see any workarounds for avoiding the duplicate code without further worsening the problem. In some languages, I could write something like this:

while ((the_input=raw_input("what to print?\n")) != "quit") {
    print the_input
}

This is definitely not Pythonic, and Python doesn’t even allow for assignment within loop conditions AFAIK.

This valid code fixes the redundancy,

while 1:
    the_input = raw_input("what to print?\n")
    if the_input == "quit":
        break
    print the_input

But doesn’t feel quite right either. The while 1 implies that this loop will run forever; I’m using a loop, but giving it a fake condition and putting the real one inside it.

Am I being too picky? Is there a better way to do this? Perhaps there’s some language construct designed for this that I don’t know of?

Answer 1

Think iterators — e.g., in this specific case:

for the_input in iter(lambda: raw_input('what to print?\n'), 'quit'):
    print the_input

Most loops in Python, except at the very lowest levels of abstractions, are best implemented as for loops with the help of some underling iterator which captures the “looping logic” — the iter built-in can help (like here), sometimes genexps (generator expressions) can, sometimes the standard library module itertools comes to the rescue.

Most often you will choose to code custom generator functions (ones using yield), or more occasionally (when you need really sophisticated state management) a custom iterator class (one defining the __iter__ special method as return self, and next [[or __next__ in the latest versions of Python]] to return “the next value from the iteration).

Capturing the looping logic apart from whatever it is that you do on the various items sequentially produced by the loop itself is the key abstraction-helper here!