7

I have read my materials, which tell that a Python iterator must have both __iter__ and __next__ methods, but an iterable just needs __iter__.

I check a list and find it has no __next__ method. When using iter() on it, it will become an iterator. Does this mean that iter() will add a __next__ method to a list to convert it to an iterator?

If yes, how does this happen?

mkrieger1
  • 19,194
  • 5
  • 54
  • 65
shuairenqin
  • 103
  • 1
  • 8

1 Answers1

6

No. iter returns an iterator, it does not convert the list into an iterator. It doesn't modify the list at all, and certainly, the list does not get a __next__ method.

>>> x = [1,2]
>>> it = iter(x)
>>> it
<list_iterator object at 0x101c021d0>
>>> x.__next__
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'list' object has no attribute '__next__'
>>>

Lists are iterables, not iterators. They implement a __iter__ method, thus they are iterable:

>>> x.__iter__
<method-wrapper '__iter__' of list object at 0x101bcf248>

But not __next__, thus they are not iterators:

>>> next(x)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'list' object is not an iterator

Iterators themselves are iterable, by definition, since they implement __iter__ as well. Consider:

>>> x = [1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> it = iter(x)
>>> it
<list_iterator object at 0x101c02358>
>>> it.__iter__
<method-wrapper '__iter__' of list_iterator object at 0x101c02358>

Most iterators should simply return themselves when you use iter on them:

>>> it2 = iter(it)
>>> it, it2
(<list_iterator object at 0x101c02358>, <list_iterator object at 0x101c02358>)
>>> it is it2
True
>>>

Indeed, this is a requirement of the iterator protocol:

"Iterators are required to have an __iter__() method that returns the iterator object itself so every iterator is also iterable and may be used in most places where other iterables are accepted."

Note, again, they are the same iterator:

>>> next(it)
1
>>> next(it2)
2
>>> next(it)
3
>>> next(it)
4
>>> next(it2)
5
>>> list(it)
[6, 7, 8, 9]
>>> next(it2)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
StopIteration

So an iterator implements __iter__ and __next__, an iterable just means that it implements __iter__. What is returned by __iter__ is an iterator, so that must implement __next__.

juanpa.arrivillaga
  • 88,713
  • 10
  • 131
  • 172
  • crucially, the `__next__` method is _on the thing_ that `iter` returns – Eevee Aug 02 '17 at 10:54
  • Note: *All* iterators must have a `__iter__` implementation that returns the iterator object itself, not just "most". This is [a requirement of the iterator protocol](https://docs.python.org/3/glossary.html#term-iterator): "Iterators are required to have an __iter__() method that returns the iterator object itself so every iterator is also iterable and may be used in most places where other iterables are accepted." Other parts of Python rely on this; one canonical and efficient duck-typing approach to test if something is an iterator is to test `obj is iter(someobj)`. – ShadowRanger Aug 02 '17 at 21:50
  • @ShadowRanger yes, you are absolutely correct. Of course, there is nothing stopping you from breaking this, but I certainly wouldn't suggest doing that. I will update to reflect this. – juanpa.arrivillaga Aug 02 '17 at 21:53