How to transpose a 3D matrix?
19,891
Solution 1
You can pass the desired dimensions to the function np.transpose
using in your case np.transpose(x_test, (0, 2, 1))
.
For example,
import numpy as np
x_test = np.arange(30).reshape(3, 2, 5)
print(x_test)
print(x_test.shape)
This will print
[[[ 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]]]
(3, 2, 5)
Now, you can transpose the matrix with the command from above
y = np.transpose(x_test, (0, 2, 1))
print(y)
print(y.shape)
which will give
[[[ 0 5]
[ 1 6]
[ 2 7]
[ 3 8]
[ 4 9]]
[[10 15]
[11 16]
[12 17]
[13 18]
[14 19]]
[[20 25]
[21 26]
[22 27]
[23 28]
[24 29]]]
(3, 5, 2)
Solution 2
Apart from transpose
(see @Cleb's answer) there are also swapaxes
and moveaxis
:
import numpy as np
mock = np.arange(30).reshape(2,3,5)
mock.swapaxes(1,2)
# array([[[ 0, 5, 10],
[ 1, 6, 11],
[ 2, 7, 12],
[ 3, 8, 13],
[ 4, 9, 14]],
[[15, 20, 25],
[16, 21, 26],
[17, 22, 27],
[18, 23, 28],
[19, 24, 29]]])
np.moveaxis(mock,2,1)
# array([[[ 0, 5, 10],
[ 1, 6, 11],
[ 2, 7, 12],
[ 3, 8, 13],
[ 4, 9, 14]],
[[15, 20, 25],
[16, 21, 26],
[17, 22, 27],
[18, 23, 28],
[19, 24, 29]]])
Author by
Ajees
I am a research scholar from the Department of Computer Science, CUSAT. Interested in areas like machine learning, NLP, Data mining, etc.
Updated on July 17, 2022Comments
-
Ajees almost 2 years
I have a 3D matrix
x_test
of size(100, 33, 66)
and I want to change its dimensions to(100, 66, 33)
.What is the most efficient way to do this using python3.5? I look for something along those lines:
y = x_test.transpose()