Iterators

In this basic post we'll explain the basics of iterators.

A Python Iterator

• If a python object contains some collection of objects, we need a way to get these object, one at a time.
• Note that:
  • A specific order is optional, but we must make sure that:
  • ..we get each object ONCE, and..
  • ..eventually we get ALL objects.
• Python iterators implement the special method next
• Each call to this special method (or a call to the next() built-in function which is the same thing) will return another object from the collection of objects.

An Iterable Object

• An iterable object, implements the iter special method
• If you call this special method (or the built-in iter() function which does the same thing), the iterable object will return an Iterator (that implements the next special method).
• Here's an example, base on the range object:

 1>>> r1 = range(5)
 2>>> it = r1.__iter__()
 3>>> it
 4<range_iterator object at 0x000001E3E5A2A0B0>
 5>>> it.__next__()
 60
 7>>> it.__next__()
 81
 9>>> it.__next__()
102
11>>> it.__next__()
123
13>>> it.__next__()
144
15>>> it.__next__()
16Traceback (most recent call last):
17  File "<stdin>", line 1, in <module>
18StopIteration
19>>>):

Creating our own Iterable and Iterator

• We'll just implement everything needed:
  • Class Letters is an iterable, so it implements the iter special method
  • When called, iter returns an instance of LetterIter, which is an iterator (it implements next)

 1class LetterIter:
 2    def __init__(self, the_letters):
 3        self.letters = the_letters
 4        self.current = -1
 5
 6    def __next__(self):
 7        self.current += 1
 8        if self.current == len(self.letters):
 9            raise StopIteration
10        return self.letters[self.current]
11    
12class Letters:
13    def __init__(self):
14        self.letters = ['a','b','c','d','e']
15
16    def __iter__(self):
17        return LetterIter(self.letters)
18    
19
20
21l1 = Letters()
22for l in l1:
23    print(l)

• Creating iterators involves writing all those methods, but there is a better way using generators.
  We'll cover generators in another post.