# Comparing two dictionaries with numpy matrices as values

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:

• Comparing two dictionaries in Python
• Comparing/combining two dictionaries

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.