python判断两个矩阵相等_将两个字典与numpy矩阵作为值进行比较

I want to assert that two Python dictionaries are equal (that means: equal amount of keys, and each mapping from key to value is equal; order is not important). A simple way would be assert A==B, however, this does not work if the values of the dictionaries are numpy arrays. How can I write a function to check in general if two dictionaries are equal?

>>> import numpy as np

>>> A = {1: np.identity(5)}

>>> B = {1: np.identity(5) + np.ones([5,5])}

>>> A == B

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

EDIT I am aware that numpy matrices shall be checked for equality with .all(). What I am looking for is a general way to check for this, without having to check isinstance(np.ndarray). Would this be possible?

Related topics without numpy arrays:

解决方案

Consider this code

>>> import numpy as np

>>> np.identity(5)

array([[ 1., 0., 0., 0., 0.],

[ 0., 1., 0., 0., 0.],

[ 0., 0., 1., 0., 0.],

[ 0., 0., 0., 1., 0.],

[ 0., 0., 0., 0., 1.]])

>>> np.identity(5)+np.ones([5,5])

array([[ 2., 1., 1., 1., 1.],

[ 1., 2., 1., 1., 1.],

[ 1., 1., 2., 1., 1.],

[ 1., 1., 1., 2., 1.],

[ 1., 1., 1., 1., 2.]])

>>> np.identity(5) == np.identity(5)+np.ones([5,5])

array([[False, False, False, False, False],

[False, False, False, False, False],

[False, False, False, False, False],

[False, False, False, False, False],

[False, False, False, False, False]], dtype=bool)

>>>

Note the the result of the comparison is a matrix, not a boolean value. Dict comparisons will compare values using the values cmp methods, which means that when comparing matrix values, the dict comparison will get a composite result. What you want to do is use

numpy.all to collapse the composite array result into a scalar boolean result

>>> np.all(np.identity(5) == np.identity(5)+np.ones([5,5]))

False

>>> np.all(np.identity(5) == np.identity(5))

True

>>>

You would need to write your own function to compare these dictionaries, testing value types to see if they are matricies, and then comparing using numpy.all, otherwise using ==. Of course, you can always get fancy and start subclassing dict and overloading cmp if you want too.