How to implement dimension exchange in Numpy

Tips for swapping dimensions in numpy

Introduction:
numpy is a powerful Python library mainly used for scientific computing and data analysis. In numpy, we often need to deal with multi-dimensional arrays, and dimension exchange of arrays is also one of the common operations. This article will introduce some techniques for swapping dimensions in numpy and provide specific code examples.

1. Dimension exchange function in numpy
In numpy, we can use the transpose() function and swapaxes() function to perform dimension exchange.

1. transpose() function
   The transpose() function is used to swap dimensions of an array, which can be achieved by specifying the order of the axes. The function prototype is:

numpy.transpose(arr, axes)

where arr is the array to be transposed, axes is the order of the axes, which can be an integer or a sequence of integers . If axes is an integer, returns a new array with dimensions swapped along that axis; if axes is a sequence of integers, returns a new array in the specified order.

2. swapaxes() function
   swapaxes() function is used to swap the two axes of the array. Its function prototype is:

numpy.swapaxes(arr, axis1 , axis2)

Among them, arr is the array of axes to be exchanged, and axis1 and axis2 are the axes to be exchanged. The swapaxes() function returns a new array whose axes are a copy of the axes of the original array, but axis1 and axis2 are swapped.

2. Examples of dimension exchange in numpy

Below we use some specific examples to demonstrate the skills of dimension exchange in numpy.

Example 1: Using the transpose() function for dimension exchange
Suppose we have a three-dimensional array with a shape of (3, 4, 2), and we want to exchange its first and second dimensions. . The code is as follows:

import numpy as np

arr = np.arange(24).reshape(3, 4, 2)
print("Original array: ")
print(arr)

new_arr = np.transpose(arr, (1, 0, 2))
print("Array after exchange:")
print(new_arr)

The running results are as follows:

Original array:
[[[ 0 1]
[ 2 3]
[ 4 5]
[ 6 7]]

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

[[16 17]
[18 19]
[20 21]
[22 23]]]

Array after exchange:
[[[ 0 1]
[ 8 9]
[16 17]]

[[ 2 3]
[10 11]
[18 19]]

[[ 4 5]
[12 13]
[20 21]]

[[ 6 7]
[14 15]
[22 23]]]

Example 2: Using the swapaxes() function for dimension exchange
Assume we have A three-dimensional array of shape (2, 5, 3), we want to swap its first and second dimensions. The code is as follows:

import numpy as np

arr = np.arange(30).reshape(2, 5, 3)
print("Original array: ")
print(arr)

new_arr = np.swapaxes(arr, 0, 1)
print("Array after swapping:")
print(new_arr)

Run result As follows:

Original array:
[[[ 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]]]

Array after exchange:
[[[ 0 1 2]
[15 16 17]]

[[ 3 4 5]
[18 19 20] ]

[[ 6 7 8]
[21 22 23]]

[[ 9 10 11]
[24 25 26]]

[ [12 13 14]
[27 28 29]]]

We demonstrated the technique of dimension exchange in numpy through the above two examples. Use the transpose() function and swapaxes() function to easily swap dimensions of arrays to meet the needs of different problems. Different dimension exchange operations can be implemented by adjusting parameters, allowing us to process multi-dimensional array data more flexibly.

The above is the detailed content of How to implement dimension exchange in Numpy. For more information, please follow other related articles on the PHP Chinese website!