I'm voting AGAINST this. order in dict has never been guaranteed. This may be changed in future. Use https://docs.python.org/3.7/library/collections.html#collections.OrderedDict if you want.

@socketpair order is guaranteed in 3.7. And @rhettinger is proposing to deprecate OrderedDict altogether.

Done. Thanks methane for your explanation.

I feel NEWS entry is too short.
And did you saw this comment ?

Oh sorry, I did't saw it.

That solution is great, and I am going to remove the tests in this PR.

And I am wondering is there necessary to keep the "test_copy_ordering" and 'test_comprehension_ordering' here? Since there have no tests for them.

I'm OK for document body, but I think someone can write more good commit message and NEWS entry. (Sorry, I'm not good English writer)

This PR updates only tutorial. But I think other commits will be made before 3.7 final can
share same NEWS entry. (Lot's of redundant NEWS entry make chengelog less usable)

Since we are editing this part, I have a question why do we use list(dict.keys()) and sorted(dict.keys())? Isn't list(dict) or sorted(dict) enough? I have seen many codes using this redundant keys call, even in Python2 testing membership.

Thanks zhangyangyu and merwok, I have made a update as your suggestions.

It's just my question. I am not sure which is actually wanted. But at least if you change it, the footprint link is not needed either.

I think that people may use d.keys() for membership or iteration because they may not be fully comfortable with the fact that iteration/membership on dicts uses keys, not items.

I don’t know if the keys method pre-dates the iteration protocol, but given that keys was never deprecated, to me it’s an indication that it’s fine if people use it, even if it’s technically redundant.

Here in the tutorial, I would go for not changing it.

I think that people may use d.keys() for membership or iteration because they may not be fully comfortable with the fact that iteration/membership on dicts uses keys, not items.

But tutorial should teach preferred way, not redundant way?

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

And if you don't make the requested changes, you will be poked with soft cushions!

I have made the requested changes; please review again

Thanks for making the requested changes!

@zhangyangyu, @JulienPalard, @merwok: please review the changes made to this pull request.

Document somewhere that two dicts with same contents, but different order of key-value pairs are equal.

i.e. assert {1:2, 3:4} == {3:4, 1:2}.

It is not obvious, and I needed to check that before use.

@socketpair Dict equality is already documented to not rely on order.

https://docs.python.org/3/library/json.html#json.load

@socketpair See #5001

Let's merge this as-is. We discussed long enough.
We can improve document later.

Thanks @shangdahao for the PR, and @methane for merging it 🌮🎉.. I'm working now to backport this PR to: 3.7.
🐍🍒⛏🤖

GH-6368 is a backport of this pull request to the 3.7 branch.