Numpy中转置transpose、T和swapaxes

来源:

https://blog.csdn.net/qq_18989901/article/details/73142472

利用Python进行数据分析时，Numpy是最常用的库，经常用来对数组、矩阵等进行转置等，有时候用来做数据的存储。

在numpy中，转置transpose和轴对换是很基本的操作，下面分别详细讲述一下，以免自己忘记。

In [1]: import numpy as np

In [2]: arr=np.arange(16).reshape(2,2,4)

In [3]: arr
Out[3]:
array([[[ 0, 1, 2, 3],
[ 4, 5, 6, 7]],

[[ 8, 9, 10, 11],
[12, 13, 14, 15]]])
如上图所示，将0-15放在一个2 2 4 的矩阵当中，得到结果如上。
现在要进行装置transpose操作，比如

In [4]: arr.transpose(1,0,2)
Out[4]:
array([[[ 0, 1, 2, 3],
[ 8, 9, 10, 11]],

[[ 4, 5, 6, 7],
[12, 13, 14, 15]]])
结果是如何得到的呢？
每一个元素都分析一下，0位置在[0,0,0]，转置为[1,0,2]，相当于把原来位置在[0,1,2]的转置到[1,0,2]，对0来说，位置转置后为[0,0,0]，同理，对1 [0,0,1]来说，转置后为[0,0，1]，同理我们写出所有如下：

其中第一列是值，第二列是转置前位置，第三列是转置后，看到转置后位置，再看如上的结果，是不是就豁然开朗了？

0 [0,0,0] [0,0,0]
1 [0,0,1] [0,0,1]
2 [0,0,2] [0,0,2]
3 [0,0,3] [0,0,3]
4 [0,1,0] [1,0,0]
5 [0,1,1] [1,0,1]
6 [0,1,2] [1,0,2]
7 [0,1,3] [1,0,3]
8 [1,0,0] [0,1,0]
9 [1,0,1] [0,1,1]
10 [1,0,2] [0,1,2]
11 [1,0,3] [0,1,3]
12 [1,1,0] [1,1,0]
13 [1,1,1] [1,1,1]
14 [1,1,2] [1,1,2]
15 [1,1,3] [1,1,3]
再看另一个结果：
In [20]: arr.T
Out[20]:
array([[[ 0, 8],
[ 4, 12]],

[[ 1, 9],
[ 5, 13]],

[[ 2, 10],
[ 6, 14]],

[[ 3, 11],
[ 7, 15]]])

In [21]: arr.transpose(2,1,0)
Out[21]:
array([[[ 0, 8],
[ 4, 12]],

[[ 1, 9],
[ 5, 13]],

[[ 2, 10],
[ 6, 14]],

[[ 3, 11],
[ 7, 15]]])
再对比转置前后的图看一下：
0 [0,0,0] [0,0,0]
1 [0,0,1] [1,0,0]
2 [0,0,2] [2,0,0]
3 [0,0,3] [3,0,0]
4 [0,1,0] [0,1,0]
5 [0,1,1] [1,1,0]
6 [0,1,2] [2,1,0]
7 [0,1,3] [3,1,0]
8 [1,0,0] [0,0,1]
9 [1,0,1] [1,0,1]
10 [1,0,2] [2,0,1]
11 [1,0,3] [3,0,1]
12 [1,1,0] [0,1,1]
13 [1,1,1] [1,1,1]
14 [1,1,2] [2,1,1]
15 [1,1,3] [3,1,1]
瞬间就明白转置了吧！其实只要动手写写，都很容易明白的。另外T其实就是把顺序全部颠倒过来，如下：
In [22]: arr3=np.arange(16).reshape(2,2,2,2)

In [23]: arr3
Out[23]:
array([[[[ 0, 1],
[ 2, 3]],

[[ 4, 5],
[ 6, 7]]],


[[[ 8, 9],
[10, 11]],

[[12, 13],
[14, 15]]]])

In [24]: arr3.T
Out[24]:
array([[[[ 0, 8],
[ 4, 12]],

[[ 2, 10],
[ 6, 14]]],


[[[ 1, 9],
[ 5, 13]],

[[ 3, 11],
[ 7, 15]]]])

In [25]: arr3.transpose(3,2,1,0)
Out[25]:
array([[[[ 0, 8],
[ 4, 12]],

[[ 2, 10],
[ 6, 14]]],


[[[ 1, 9],
[ 5, 13]],

[[ 3, 11],
[ 7, 15]]]])

转置就是这样子，具体上面aar3转置前后的位置，就不写了。
下面说说swapaxes，轴对称。

话不多，上结果

In [27]: arr.swapaxes(1,2)
Out[27]:
array([[[ 0, 4],
[ 1, 5],
[ 2, 6],
[ 3, 7]],

[[ 8, 12],
[ 9, 13],
[10, 14],
[11, 15]]])

In [28]: arr.transpose(0,2,1)
Out[28]:
array([[[ 0, 4],
[ 1, 5],
[ 2, 6],
[ 3, 7]],

[[ 8, 12],
[ 9, 13],
[10, 14],
[11, 15]]])
发现了吧，其实swapaxes其实就是把矩阵中某两个轴对换一下，不信再看一个：
In [29]: arr3
Out[29]:
array([[[[ 0, 1],
[ 2, 3]],

[[ 4, 5],
[ 6, 7]]],


[[[ 8, 9],
[10, 11]],

[[12, 13],
[14, 15]]]])

In [30]: arr3.swapaxes(1,3)
Out[30]:
array([[[[ 0, 4],
[ 2, 6]],

[[ 1, 5],
[ 3, 7]]],


[[[ 8, 12],
[10, 14]],

[[ 9, 13],
[11, 15]]]])

In [31]: arr3.transpose(0,3,2,1)
Out[31]:
array([[[[ 0, 4],
[ 2, 6]],

[[ 1, 5],
[ 3, 7]]],


[[[ 8, 12],
[10, 14]],

[[ 9, 13],
[11, 15]]]])