slight problem with testing problems

[pjritee]

I am fixing up the duplication in the dictionary problems and rewrote the dictionary iteration example. It's the same as the old one - find the keys whose values are bigger than another value. Given that the hashing of strings seems to be different on different OSs I decided that the function should return the sorted list. However, an interesting problem arose (at least under Linux) I wrote a solution that didn't sort before returning and the test big_keys({'a':24, 'x':30, 't':12, 'n':10}, 15) -> ['a', 'x'] works but the hidden alt test big_keys({'a':24, 'e':30, 't':12, 'n':10}, 40) -> ['a', 'e', 'n', 't']

doesn't work and I get

Make sure your code also works for similar inputs

Unfortunately this doesn't give any hints to the student as to what went wrong - it was just a fluke that the first test generated the list in sorted order itself. Any suggestions on what to do?

[sapi]

The best solution would be to just require the student to return the list in any order, but only test for unordered equality.

You could either do

self.assertEqual(sorted(expected), sorted(student_result))

or

self.assertEqual(set(expected), set(student_result))

[pjritee]

Yes - I was coming to that conclusion - thanks

[jgat]

Note - I suggest sorted() over set() for checking unordered equality, since it preserves any incorrect duplicates.

[pjritee]

Yes - I did that (for the same reason)