I have a just a short question about the attributes shape and ndim. How can an array have more dimensions than two, but matrices are limited to a n x m shape? Would it not be necessary to have shapes with something like a m x n x o shape for a three-dimensional array?
Best regards
I think you are confusing things here, np.array can have any number of dimensions and np.matrix only two, np.matrix is the specific case where ndim=2 and have special matrix operators. From NumPy documentation of np.matrix:
Note:
It is no longer recommended to use this class, even for linear algebra.
Instead, use regular arrays. The class may be removed in the future.
Returns a matrix from an array-like object, or from a string of data.
A matrix is a specialized 2-D array that retains its 2-D nature through operations.
It has certain special operators, such as * (matrix multiplication) and ** (matrix power).
Also, notice this example so you can see the difference:
a = np.matrix([[2,2],[2,2]])
print(a**2)
This would return:
matrix([[8, 8],
[8, 8]])
Because matrix "a" squared is that, but if you do the same thing with array:
a = np.array([[2,2],[2,2]])
print(a**2)
It will return:
array([[4, 4],
[4, 4]])
Because it applied the square in all elements if you want the behaviour of np.matrix you would have to use np.dot
I think they will probably remove np.matrix in future versions because it doesn't add anything really useful other than bugging the mind of developers.
