For two dictionaries d1 and d2 defined as
d1 = {'foo':123, 'bar':789}
d2 = {'bar':789, 'foo':123}
The order of the keys is preserved in Python 3.6+. This is evident when we loop through the dictionary and print the items.
>>> for x in d1.items():
... print(x)
...
('foo', 123)
('bar', 789)
>>> for x in d2.items():
... print(x)
...
('bar', 789)
('foo', 123)
Why does Python still consider d1 and d2 to be equal?
>>> d1 == d2
True
Dictionaries are hash tables, order is not supposed to matter. In python 3.6+ dictionaries are in insertion order but that is just how they are implemented. Order doesn't matter for equality. If you want order to matter in equality use an OrderedDict.
from collections import OrderedDict
d1 = OrderedDict({'foo':123, 'bar':789})
d2 = OrderedDict({'bar':789, 'foo':123})
print(d1 == d2) # False
dicts.. The fact that order of insertion is maintained can again change in any future version of Python
dict maintains insertion order.
Dictionaries were unordered up until Python 3.6. Therefore the only sensible way to test for equality was to ignore the order. When Python 3.6 ordered dictionaries this was an implementation detail. Since Python 3.7 the insertion order can be relied upon. But changing the behavior to only consider dictionaries only equal if the keys are in the same order would break a whole lot of code. And quite frankly I think it is more useful to compare dictionaries without taking the ordering into account.
OrderedDict-OrderedDict([("foo", 123), ("bar", 789)]) != OrderedDict([("bar", 789), ("foo", 123)]).