numpy中的数组拆分

垂直拆分：numpy.vsplit(数组，份数)->(数组片段)

In [20]: c = np.arange(1,13).reshape(6,2)

In [21]: c
Out[21]: 
array([[ 1,  2],
       [ 3,  4],
       [ 5,  6],
       [ 7,  8],
       [ 9, 10],
       [11, 12]])
       
In [22]: np.vsplit(c,3)
Out[22]: 
[array([[1, 2],
        [3, 4]]), array([[5, 6],
        [7, 8]]), array([[ 9, 10],
        [11, 12]])]

水平拆分：numpy.hsplit(数组，份数)->(数组片段)

In [27]: d =c.T

In [28]: d
Out[28]: 
array([[ 1,  3,  5,  7,  9, 11],
       [ 2,  4,  6,  8, 10, 12]])

In [29]: np.hsplit(d,3)
Out[29]: 
[array([[1, 3],
        [2, 4]]), array([[5, 7],
        [6, 8]]), array([[ 9, 11],
        [10, 12]])]

numpy.dsplit(数组，份数)->(数组片段)

In [32]: e = np.dstack((a,b))

In [33]: e
Out[33]: 
array([[[11, 21],
        [12, 22],
        [13, 23]],

       [[14, 24],
        [15, 25],
        [16, 26]],

       [[17, 27],
        [18, 28],
        [19, 29]]])

In [35]: np.dsplit(e,2)
Out[35]: 
[array([[[11],
         [12],
         [13]],
 
        [[14],
         [15],
         [16]],
 
        [[17],
         [18],
         [19]]]), array([[[21],
         [22],
         [23]],
 
        [[24],
         [25],
         [26]],
 
        [[27],
         [28],
         [29]]])]

接下来我们研究一下e的转置矩阵

In [42]: np.shape(f)
Out[42]: (2, 3, 3, 1)

In [44]: f[0]
Out[44]: 
array([[[11],
        [12],
        [13]],

       [[14],
        [15],
        [16]],

       [[17],
        [18],
        [19]]])

In [45]: f[0].T
Out[45]: 
array([[[11, 14, 17],
        [12, 15, 18],
        [13, 16, 19]]])

In [46]: f[0].T[0]
Out[46]: 
array([[11, 14, 17],
       [12, 15, 18],
       [13, 16, 19]])

In [47]: f[0].T[0].T
Out[47]: 
array([[11, 12, 13],
       [14, 15, 16],
       [17, 18, 19]])