与输入所有数字不同，我将创建一个（5，8）数组，如下所示：

In [51]: arr = np.arange(40).reshape(5,8)

In [52]: arr
Out[52]: 
array([[ 0,  1,  2,  3,  4,  5,  6,  7],
       [ 8,  9, 10, 11, 12, 13, 14, 15],
       [16, 17, 18, 19, 20, 21, 22, 23],
       [24, 25, 26, 27, 28, 29, 30, 31],
       [32, 33, 34, 35, 36, 37, 38, 39]])

有很多种方法可以将数组分成块，然后重新排列它们。一种直接的方法是使用np.block并连接4个两列块。这可能不是最通用的，但很容易写出：

In [53]: np.block([[arr[:,0:2],arr[:,4:6]],[arr[:,2:4],arr[:, 6:8]]])
Out[53]: 
array([[ 0,  1,  4,  5],
       [ 8,  9, 12, 13],
       [16, 17, 20, 21],
       [24, 25, 28, 29],
       [32, 33, 36, 37],
       [ 2,  3,  6,  7],
       [10, 11, 14, 15],
       [18, 19, 22, 23],
       [26, 27, 30, 31],
       [34, 35, 38, 39]])

split也可以将其划分为以下4个块：

In [54]: alist = np.split(arr, 4, axis=1)

In [55]: alist
Out[55]: 
[array([[ 0,  1],
        [ 8,  9],
        [16, 17],
        [24, 25],
        [32, 33]]),
 array([[ 2,  3],
        [10, 11],
        [18, 19],
        [26, 27],
        [34, 35]]),
 array([[ 4,  5],
        [12, 13],
        [20, 21],
        [28, 29],
        [36, 37]]),
 array([[ 6,  7],
        [14, 15],
        [22, 23],
        [30, 31],
        [38, 39]])]

下面是一个工作的整形和转置序列：

In [64]: arr.reshape(5,2,2,2).transpose(2,0,1,3).reshape(10,4)
Out[64]: 
array([[ 0,  1,  4,  5],
       [ 8,  9, 12, 13],
       [16, 17, 20, 21],
       [24, 25, 28, 29],
       [32, 33, 36, 37],
       [ 2,  3,  6,  7],
       [10, 11, 14, 15],
       [18, 19, 22, 23],
       [26, 27, 30, 31],
       [34, 35, 38, 39]])

我经过一些反复试验，在以前这种重组经验的帮助下，达到了这个目的。