How do I consistently flatten a numpy array?

问题

from numpy import array, eye, matrix

x = array([1, 0])
A = eye(2)
print(A.dot(x))

prints [1. 0.].

On the other hand,

B = matrix([[1, 0], [0, 1]])
print(B.dot(x))

prints [[1 0]] which is a 1-by-2 array. Furthermore,

print(B.dot(x).flatten())

also prints [[1 0]].

This is rather annoying. Why does flatten fail here and how else can I get this into the 1-d shape?

回答1:

Stop using matrix. numpy.matrix.flatten returns a 1-row matrix, because that's as flat as matrix instances get. If for some reason you are dead set on using matrix, convert to ndarray with matrix.A before flattening:

flat = whatever_matrix.A.flatten()

or just use A1 to get a flat ndarray directly:

flat = whatever_matrix.A1

来源：https://stackoverflow.com/questions/50453626/how-do-i-consistently-flatten-a-numpy-array