Description of issue or feature request:
In #1783 we dropped support for Python 3.6 and inside this pr, we decided to drop OrderedDicts as well: e3b267e.
The argument was that inside python 3.7+ dictionaries insertion order is preserved: https://docs.python.org/3/whatsnew/3.7.html.
Even though the insertion order is preserved we should keep in mind that there are still differences between OrderedDicts and regular dictionaries like when comparing them:
>>>OrderedDict([(1,1), (2,2)]) ==OrderedDict([(2,2), (1,1)]) False>>>dict([(1,1), (2,2)]) ==dict([(2,2), (1,1)]) True
The question is:
Is it a problem if we have two dictionaries with data in a different order that are not the same?
If yes we have two solutions:
- revert e3b267e
- inside the
__eq__ implementations make sure this will happen:
>>>list(dict([(1,1), (2,2)]).keys()) ==list(dict([(2,2), (1,1)]).keys())
False
or in other words where we are using dictionaries make sure when comparing them that the order makes a difference.
Description of issue or feature request:
In #1783 we dropped support for Python 3.6 and inside this pr, we decided to drop
OrderedDicts as well: e3b267e.The argument was that inside python 3.7+ dictionaries insertion order is preserved: https://docs.python.org/3/whatsnew/3.7.html.
Even though the insertion order is preserved we should keep in mind that there are still differences between
OrderedDicts and regular dictionaries like when comparing them:The question is:
Is it a problem if we have two dictionaries with data in a different order that are not the same?
If yes we have two solutions:
__eq__implementations make sure this will happen:or in other words where we are using dictionaries make sure when comparing them that the order makes a difference.