Best way to declare numpy.array equality?

Question

I want to do some unit tests for my application, and I need to compare two arrays. Since array.__eq__ returns a new array (so TestCase.assertEqual fails), what is the best way to assert equality?

I am currently using

 self.assertTrue((arr1 == arr2).all()) 

but i don't really like it

Answer 1

check assert functions in numpy.testing e.g.

assert_array_equal

for floating point arrays, equality checking may fail and assert_almost_equal more reliable.

Refresh

A few versions ago, numpy got assert_allclose which is now my favorite, because it allows us to indicate both absolute and relative error and does not require decimal rounding as a criterion for proximity.

Answer 2

I think (arr1 == arr2).all() looks pretty pretty. But you can use:

 numpy.allclose(arr1, arr2) 

but it’s not exactly the same.

An alternative is almost the same as your example:

 numpy.alltrue(arr1 == arr2) 

Note that scipy.array is actually a numpy.array reference. This makes it easier to find documentation.

Answer 3

I believe that using self.assertEqual(arr1.tolist(), arr2.tolist()) is the easiest way to compare arrays with unittest.

I agree that this is not the best solution, and it is probably not the fastest, but probably more uniform with the rest of your test cases, you get the whole unittest error description and it is really easy to implement.

Answer 4

With Python 3.2, you can use assertSequenceEqual(array1.tolist(), array2.tolist()) .

This adds a value showing the exact elements in which the arrays differ.

Answer 5

In my tests, I use this:

 try: numpy.testing.assert_array_equal(arr1, arr2) res = True except AssertionError as err: res = False print (err) self.assertTrue(res) 

Answer 6

np.linalg.norm(arr1 - arr2) < 1e-6