python 多维数组轴_python – numpy通过任意轴重塑多维数组

每个数组都有一个自然(1D扁平)的元素.重塑数组时,就好像它首先被展平(从而获得自然顺序),然​​后重新整形：

In [54]: z.ravel()

Out[54]:

array([ 0, 3, 6, 1, 4, 7, 2, 5, 8, 9, 12, 15, 10, 13, 16, 11, 14,

17])

In [55]: z.ravel().reshape(2*3, 3)

Out[55]:

array([[ 0, 3, 6],

[ 1, 4, 7],

[ 2, 5, 8],

[ 9, 12, 15],

[10, 13, 16],

[11, 14, 17]])

请注意,在“自然顺序”中,0和1相距很远.无论你重塑它,0和1都不会沿着最后一个轴彼此相邻,这就是你想要的所需数组：

desired = np.array([[ 0, 1, 2],

[ 3, 4, 5],

[ 6, 7, 8],

[ 9, 10, 11],

[12, 13, 14],

[15, 16, 17]])

这需要一些重新排序,在这种情况下可以通过交换：

In [53]: z.swapaxes(1,2).reshape(2*3, 3)

Out[53]:

array([[ 0, 1, 2],

[ 3, 4, 5],

[ 6, 7, 8],

[ 9, 10, 11],

[12, 13, 14],

[15, 16, 17]])

因为swapaxes(1,2)将值放在所需的顺序中

In [56]: z.swapaxes(1,2).ravel()

Out[56]:

array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,

17])

In [57]: desired.ravel()

Out[57]:

array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,

17])

请注意,重塑方法还有一个顺序参数,可用于控制(C-或F-)顺序,从数组中读取元素并将其放置在重新形成的数组中.但是,我不认为这对你的情况有帮助.

考虑重塑极限的另一种方法是说所有重塑后跟ravel是相同的：

In [71]: z.reshape(3,3,2).ravel()

Out[71]:

array([ 0, 3, 6, 1, 4, 7, 2, 5, 8, 9, 12, 15, 10, 13, 16, 11, 14,

17])

In [72]: z.reshape(3,2,3).ravel()

Out[72]:

array([ 0, 3, 6, 1, 4, 7, 2, 5, 8, 9, 12, 15, 10, 13, 16, 11, 14,

17])

In [73]: z.reshape(3*2,3).ravel()

Out[73]:

array([ 0, 3, 6, 1, 4, 7, 2, 5, 8, 9, 12, 15, 10, 13, 16, 11, 14,

17])

In [74]: z.reshape(3*3,2).ravel()

Out[74]:

array([ 0, 3, 6, 1, 4, 7, 2, 5, 8, 9, 12, 15, 10, 13, 16, 11, 14,

17])

因此,如果所需数组的ravel不同,则无法获得它只是重新整形.

对于order =’F’进行重新整形也是如此,前提是你还要使用order =’F’进行调整：

In [109]: z.reshape(2,3,3, order='F').ravel(order='F')

Out[109]:

array([ 0, 9, 1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7, 16, 8,

17])

In [110]: z.reshape(2*3*3, order='F').ravel(order='F')

Out[110]:

array([ 0, 9, 1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7, 16, 8,

17])

In [111]: z.reshape(2*3,3, order='F').ravel(order='F')

Out[111]:

array([ 0, 9, 1, 10, 2, 11, 3, 12, 4, 13, 5, 14, 6, 15, 7, 16, 8,

17])

可以使用两个重塑获得所需的数组：

In [83]: z.reshape(2, 3*3, order='F').reshape(2*3, 3)

Out[83]:

array([[ 0, 1, 2],

[ 3, 4, 5],

[ 6, 7, 8],

[ 9, 10, 11],

[12, 13, 14],

[15, 16, 17]])

但是我偶然发现了这件事.

如果我完全误解了你的问题并且x和y是给号(而不是z)那么你可以使用row_stack而不是dstack获得所需的数组：

In [88]: z = np.row_stack([x, y])

In [89]: z

Out[89]:

array([[ 0, 1, 2],

[ 3, 4, 5],

[ 6, 7, 8],

[ 9, 10, 11],

[12, 13, 14],

[15, 16, 17]])