利用Python進行數據分析時,Numpy是最常用的庫,經常用來對數組、矩陣等進行轉置等,有時候用來做數據的存儲。
在numpy中,轉置transpose和軸對換是很基本的操作,下面分別詳細講述一下,以免自己忘記。
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]]])
如上圖所示,將0-15放在一個2 2 4 的矩陣當中,得到結果如上。
現在要進行裝置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]]])
結果是如何得到的呢?
每一個元素都分析一下,0位置在[0,0,0],轉置爲[1,0,2],相當於把原來位置在[0,1,2]的轉置到[1,0,2],對0來說,位置轉置後爲[0,0,0],同理,對1 [0,0,1]來說,轉置後爲[0,0,1],同理我們寫出所有如下:
其中第一列是值,第二列是轉置前位置,第三列是轉置後,看到轉置後位置,再看如上的結果,是不是就豁然開朗了?
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]
再看另一個結果:
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]]])
再對比轉置前後的圖看一下:
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]
瞬間就明白轉置了吧!其實只要動手寫寫,都很容易明白的。另外T其實就是把順序全部顛倒過來,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]]]])
轉置就是這樣子,具體上面aar3轉置前後的位置,就不寫了。
下面說說swapaxes,軸對稱。
話不多,上結果
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]]])
發現了吧,其實swapaxes其實就是把矩陣中某兩個軸對換一下,不信再看一個:
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]]]])
哈哈,只要動手做做,會發現其實沒有那麼困難,不能只看
紙上得來終覺淺,絕知此事要躬行!共勉