Chris Angelico wrote: > On Fri, Sep 19, 2014 at 8:59 PM, Steven D'Aprano > <steve+comp.lang.pyt...@pearwood.info> wrote: >>> https://docs.python.org/3.4/library/stdtypes.html#dict-views >>> """If keys, values and items views are iterated over with no >>> intervening modifications to the dictionary, the order of items will >>> directly correspond.""" >>> So if iterating over d.keys() and then d.values() with no mutations is >>> guaranteed to give the same order, then so is iterating over d.keys(), >>> then d.keys(), then d.values(), >> >> Not so! So long as the iteration of values() matched the *second* >> iteration of keys(), it would be allowed. There's nothing which says that >> the first iteration of keys() has to match the second. > > I disagree. Between the first iteration of keys() and the iteration of > values(), there were no modifications to the dictionary - another > iteration over keys() isn't a modification - ergo the guarantee ought > to hold.
All it says is that keys/values/items will correspond, not keys/keys, etc. Hmmm. On second thought, there is a problem with this: a = list(mydict) c = list(mydict.items()) # now the internal order is shuffled b = list(mydict.values()) assert list(zip(a, b)) == c and the assertion fails. So I guess that rules out dict internal order changing during a single run of the interpreter, unless the dict is modified. Anyway, that's just dicts. Custom iterables can change any time they like. -- Steven -- https://mail.python.org/mailman/listinfo/python-list
