np.newaxis 实现为 numpy.ndarray(多维数组)增加一个轴

>> type(np.newaxis)
NoneType
>> np.newaxis == None
True

>> x = np.arange(3)
>> x
array([0, 1, 2])
>> x.shape
(3,)

>> x[:, np.newaxis]
array([[0],
    [1],
    [2]])

>> x[:, None]
array([[0],
    [1],
    [2]])

>> x[:, np.newaxis].shape
 (3, 1)

>>> X = np.array([[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]])
>>> X[:, 1]
array([2, 6, 10])    % 这里是一个行
>>> X[:, 1].shape    % X[:, 1] 的用法完全等同于一个行，而不是一个列，
(3, )

>>>X[:, 1][:, np.newaxis]
array([[2],
   [6],
   [10]])

>>>X_sub = np.hstack([X[:, 1][:, np.newaxis], X[:, 3][:, np.newaxis]])      
          % hstack：horizontal stack，水平方向上的层叠
>>>X_sub
array([[2, 4]
   [6, 8]
   [10, 12]])

>> X[:, [1, 3]]
array([[ 2, 4],
    [ 6, 8],
    [10, 12]])

>> X = np.random.randint(0, 9, (2, 3))
>> mean_X = np.mean(X, axis=0)
>> X - mean_X           # 这样做是没有问题的

>> mean_X = np.mean(X, axis=1)
>> X - mean_X
ValueError: operands could not be broadcast together with shapes (2,3) (2,)