Why is the range object "not an iterator"?

Question

I wrote this and expected 0:

>>> x = range(20)
>>> next(x)

Instead I got:

TypeError: 'range' object is not an iterator

But I thought it was a generator?

The initial answer yielded the same thing I initially said to myself: it's an iterable, not an interator. But then, that wouldn't explain why this works, if both are simply generators:

>>> x = (i for i in range(30))
>>> next(x)
0

Also relevant: http://docs.python.org/3/library/stdtypes.html#ranges — user2357112, Feb 15 '14 at 21:35

score 32 · Answer 1 · edited Mar 03 '17 at 20:12

32

The range object is iterable. However, it's not an iterator.

To get an iterator, you need to call iter() first:

>>> r=range(5,15)
>>> next(iter(r))
5
>>> next(iter(r))
5
>>> next(iter(r))
5
>>> next(iter(r))
5
>>> i=iter(r)
>>> next(i)
5
>>> next(i)
6
>>> next(i)
7
>>> next(i)
8
>>> iter(r)
<range_iterator object at 0x10b0f0630>
>>> iter(r)
<range_iterator object at 0x10b0f0750>
>>> iter(r)
<range_iterator object at 0x10b0f0c30>

Edit: But be careful not to call iter() with every call to next(). It creates a new iterator at index 0.

edited Mar 03 '17 at 20:12

Bruno Bronosky

66,273
12
162
149

answered Feb 15 '14 at 21:26

NPE

486,780
108
951
1,012

That's the same thing I told myself initially. But then, why does this work: `next( (i for i in range(30) )`, since that's a generator expression? Actually, let me add this to the question. – temporary_user_name Feb 15 '14 at 21:29
"It is indeed a generator" - no, it isn't. – user2357112 Feb 15 '14 at 21:30
Just spotted this http://stackoverflow.com/questions/13092267/if-range-is-a-generator-in-python-3-3-why-can-i-not-call-next-on-a-range?rq=1. This question needs to be marked as duplicate. – temporary_user_name Feb 15 '14 at 21:32
@user2357112: Corrected, thanks for pointing out the error. – NPE Feb 15 '14 at 21:33
1

@Aerovista: Because a generator expression evaluates to an iterator? – RemcoGerlich Feb 15 '14 at 21:35
1

@NPE tldr; generators can be treated as iterators, as I thought. So that information is unrelated-- the correct answer is that range() objects are immutable and special and are not in fact simple generators. – temporary_user_name Feb 15 '14 at 21:35
@NPE: No, you can `next` a generator. Generators are iterators. – user2357112 Feb 15 '14 at 21:39
1

@user2357112: Oh, I wasn't calling it. This explains it. Long day, I'm off to bed. – NPE Feb 15 '14 at 21:40
@Aerovistae, generators *are* iterators. A range object isn't a generator at all, simple or otherwise. – Mike Graham Feb 15 '14 at 21:48

score 13 · Accepted Answer · answered Feb 15 '14 at 21:35

13

range returns an iterable, not an iterator. It can make iterators when iteration is necessary. It is not a generator.

A generator expression evaluates to an iterator (and hence an iterable as well).

answered Feb 15 '14 at 21:35

Mike Graham

73,987
14
101
130

2

Note that you get an iterator out of this iterable by calling iter() on it. You can then call the next() builtin on the result from iter(range(...)). It's confusing but just one of those things you learn when you dive deeply enough into the language. Note that for loops implicitly call iter(expression) once and then next(results_from_iter) for every pass through the loop. – Jim Dennis Jan 16 '18 at 06:33

score 6 · Answer 3 · answered Feb 15 '14 at 21:38

The next builtin calls the __next__ hook method. So, range objects have a well defined __iter__, but not a well-defined __next__.

iterable objects have __iter__ defined, iterator objects have well defined __next__ (typically with an __iter__ method which simply returns self).

Bleeding Fingers · Answer 4 · 2014-02-15T22:04:28.147

It's because the next function calls the next method of the object that passed in.

next(...)
    x.next() -> the next value, or raise StopIteration

listiterators and generators both have the next method.

>>> iter(range(1)).__class__.next
<slot wrapper 'next' of 'listiterator' objects>
>>> iter(x for x in range(1)).__class__.next
<slot wrapper 'next' of 'generator' objects>

But a list doesn't have it. And that is the reason why it raises that exception.

>>> list.next
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: type object 'list' has no attribute 'next'

next doesn't care much about whether the object it's passed is an iterator or not.

>>> class Foo():
...     def next(self):
...             return "foo"
... 
>>> foo = Foo()
>>> next(foo)
'foo'
>>> next(foo)
'foo'

But adding the next method doesn't necessarily make it a collection/sequence/iterable.

>>> class Foo():
...     def next(self):
...             return "Foo"
>>> [x for x in Foo()]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: iteration over non-sequence
>>> iter(Foo())
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: iteration over non-sequence

But adding the __iter__ method to it makes it one.

>>> class Foo():
...     def next(self):
...             return "Foo"
...     def __iter__(self): return self
... 
>>> [x for x in Foo()]
^CTraceback (most recent call last):
  File "<stdin>", line 1, in <module>
KeyboardInterrupt
>>> iter(Foo())
<__main__.Foo instance at 0x7fd77307c488>

The next seems to have some builtin intelligence when it comes to list.

>>> class Foo():
...     pass
... 
>>> foo = Foo()
>>> next(foo)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: instance has no next() method
>>> next(range(20))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: list object is not an iterator

Why is the range object "not an iterator"?

4 Answers4

Linked