• Iterator

An iterable object is an object that implements __iter__, which is expected to return an iterator object.

An iterator object implements __next__, which is expected to return the next element of the iterable object that returned it, and to raise a StopIteration exception when no more elements are available.

In the simplest case, the iterable will implement __next__ itself and return self in __iter__. However, this has its limitations and may produce unexpected results in concurrent environments (e.g. with the multiprocessing API).

You can use iterables in for loops, to construct lists with list comprehensions, or as input arguments for the list function.

Example Iterator

Here is an iterator that returns a random number of 1's:

   1 import random
   2 
   3 class RandomIterable:
   4     def __iter__(self):
   5         return self
   6     def __next__(self):
   7         if random.choice(["go", "go", "stop"]) == "stop":
   8             raise StopIteration  # signals "the end"
   9         return 1

Q: Why is __iter__ there, if it just returns self?

A: This is a very simple case. More complex iterables may very well return separate iterator objects.

Q: When would I need an extra iterator?

A: Iterators will typically need to maintain some kind of position state information (e.g., the index of the last element returned). If the iterable maintained that state itself, it would become inherently non-reentrant (meaning you could use it only one loop at a time).

   1 for eggs in RandomIterable():
   2     print(eggs)

You can also use it in list construction:

   1 >>> list(RandomIterable())
   2 [1]
   3 >>> list(RandomIterable())
   4 []
   5 >>> list(RandomIterable())
   6 [1, 1, 1, 1, 1]
   7 >>> list(RandomIterable())
   8 [1]

...both of these uses require __iter__.

An object isn't iterable unless it provides __iter__. And for an object to be a valid iterator, it must provide __next__.

Manual usage

Although you won't need this in most cases, you can manually get the iterator from an iterable object by using the iter() function. Similary, you can manually call __next___ using the next() function.

Links

• PEP-234: Iterators

• Itertools: Functions creating iterators for efficient looping

• Functional programming and iterators

• Python iterator basics (how they work + examples)

See also: Generators