2

我想重塑 Numpy 数组 A，不是通过附加正常的下一行，而是通过附加它之后的每 N 行。

前任：

A = [[1  2  3  4]
[5  6  7  8]
[9  10 11 12]
[13 14 15 16]
[17 18 19 20]
[21 22 23 24]]

现在，我想从 A 构造一个像这样大小为 2x12 的数组

 A = [[1 2 3 4 9 10 11 12 17 18 19 10]
 [5 6 7 8 13 14 15 16 21 22 23 24]]

在这里您可以看到当前行的每第二行都附加到它并形成了新的重新整形数组。

4

2 回答 2

2

这种方法怎么样？

# extract alternative rows starting from 0th row (1st row)
In [18]: A[0::2]
Out[18]: 
array([[ 1,  2,  3,  4],
       [ 9, 10, 11, 12],
       [17, 18, 19, 20]])

# and then flatten to 1D array
In [19]: A[0::2].flatten()
Out[19]: array([ 1,  2,  3,  4,  9, 10, 11, 12, 17, 18, 19, 20])


# extract alternative rows starting from 2nd row
In [17]: A[1::2]
Out[17]: 
array([[ 5,  6,  7,  8],
       [13, 14, 15, 16],
       [21, 22, 23, 24]])


# and then flatten to 1D array
In [20]: A[1::2].flatten()
Out[20]: array([ 5,  6,  7,  8, 13, 14, 15, 16, 21, 22, 23, 24])


# to get 2D, just put them in `np.vstack` (in the order you want the final array)
In [21]: np.vstack((A[0::2].flatten(), A[1::2].flatten()))
Out[21]: 
array([[ 1,  2,  3,  4,  9, 10, 11, 12, 17, 18, 19, 20],
       [ 5,  6,  7,  8, 13, 14, 15, 16, 21, 22, 23, 24]])
于 2017-04-01T12:30:09.400 回答
2

您可以使用简单的索引，array.ravel()并且np.vstack()：

In [37]: np.vstack((A[::2].ravel(), A[1::2].ravel()))
Out[37]: 
array([[ 1,  2,  3,  4,  9, 10, 11, 12, 17, 18, 19, 20],
       [ 5,  6,  7,  8, 13, 14, 15, 16, 21, 22, 23, 24]])
于 2017-04-01T11:58:32.763 回答