homepage

win XP sp2, Py3.0c2
I had to face an annoying problem when iterating over a map object.
With a range class, this works
>>> r = range(5)
>>> list(r)
[0, 1, 2, 3, 4]
With dict_keys/values/items objects, the following works
>>> d = {1: 'a', 2:'b', 3:'c'}
>>> list(d.keys())
[1, 2, 3]
>>> list(d.values())
['a', 'b', 'c']
>>> list(d.items())
[(1, 'a'), (2, 'b'), (3, 'c')]
But with a map object...
>>> def plus(i):
 return i + 1
>>> a = list(range(3))
>>> a
[0, 1, 2]
>>> r = map(plus, a)
>>> r
<map object at 0x01371570>
>>> for e in r: print(e)
1
2
3
>>> list(r)
[]
>>> #empty list!
Bug or feature?

Feature :-)
You will get the expected result if you skip the step where you ran the
for-loop over r before running list(). Either listing or for-looping
will exhaust the iterator. This is how iterators work.

2008年11月17日 Raymond Hettinger <report@bugs.python.org>
>
> Raymond Hettinger <rhettinger@users.sourceforge.net> added the comment:
>
> Feature :-)
>
> You will get the expected result if you skip the step where you ran the
> for-loop over r before running list(). Either listing or for-looping
> will exhaust the iterator. This is how iterators work.
>
> ----------
> nosy: +rhettinger
> resolution: -> invalid
> status: open -> closed
>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <http://bugs.python.org/issue4337>
> _______________________________________
>
Thanks for the reply and sorry for the noise. Indeed, you are right and for
some "strange personal reasons" (bad habits?), I frequently fall in this
trap.
 return i + 1
[1, 2, 3, 4]
>>>
Regards

Dict views and range objects are *iterables* because they are based on
reusable information. Map, filter, and similar objects are *iterators*
because they are based on iterables that could be once-through
iterators. The built-in function entries carefully specific which is which.