NumPy diag()

import numpy as np

# create a 1D array
array1 = np.arange(3)

# create a 2D array
array2 = np.arange(9).reshape(3,3)

# create a 2D array with elements of arr1 as diagonal
diagonalArray1 = np.diag(array1)

print('Array1:\n', array1)
print('Array1 as diagonal elements:\n', diagonalArray1)

# extract diagonal elements from arr2
diagonalArray2 = np.diag(array2)


print('\nArray2\n', array2)
print('Extract diagonal of Array2\n', diagonalArray2)

'''
Array1:
[0 1 2]
Array1 as diagonal elements:
 [[0 0 0]
 [0 1 0]
 [0 0 2]]

Array2
 [[0 1 2]
 [3 4 5]
 [6 7 8]]
Extract diagonal of Array2
 [0 4 8]
'''

diag() 语法

diag() 的语法是

numpy.diag(array, k = 0)

diag() 参数

diag() 方法接受以下参数

• array - 输入数组（可以是 array_like）
• k（可选）- 一个整数，表示要检索的对角线

diag() 返回值

diag() 方法要么返回一个新 ndarray，将 1D 数组中的值作为其对角线，要么返回一个包含给定 ndarray 对角线元素的 1D 数组。

示例 1：使用 1D 数组创建对角线数组

当一个 1D 数组被传递给 diag() 时，它会创建一个以给定数组作为对角线元素的对角线数组。

如前所述，我们可以使用 k 参数来控制结果数组中对角线元素的位置。

让我们看一个例子。

import numpy as np

# create a 1D array
array1 = np.arange(1, 4)

# create a 2D array with elements of array1 as the main diagonal
mainDiagonal = np.diag(array1)

# create a 2D array with elements of array1 as diagonal above the main diagonal
upperDiagonal = np.diag(array1, k = 1)

# create a 2D array with elements of array1 as  diagonal below the main diagonal
lowerDiagonal = np.diag(array1, k = -1)

print('Array1:\n',array1)
print('Array1 as main diagonal elements:\n', mainDiagonal)
print('Array1 as diagonal elements above main diagonal:\n', upperDiagonal)
print('Array1 as diagonal elements below main diagonal:\n', lowerDiagonal)

输出

Array1:
[1 2 3]
Array1 as main diagonal elements:
 [[1 0 0]
 [0 2 0]
 [0 0 3]]
Array1 as diagonal elements above main diagonal:
 [[0 1 0 0]
 [0 0 2 0]
 [0 0 0 3]
 [0 0 0 0]]
Array1 as diagonal elements below main diagonal:
 [[0 0 0 0]
 [1 0 0 0]
 [0 2 0 0]
 [0 0 3 0]]

示例 2：从 2D 数组提取对角线

当一个 2D 数组被传递给 diag() 时，它会创建一个 1D 数组，其中包含给定数组的对角线元素。

让我们看一个例子。

import numpy as np

# create a 2D array
array1 = np.arange(1, 10).reshape(3,3)

# create a 1D array with main diagonal as elements
mainDiagonal = np.diag(array1)

# create a 1D array with diagonal elements of arr1 one step above the main diagonal
upperDiagonal = np.diag(array1, k = 1)

# create a 1D array with diagonal elements of arr1 one step below the main diagonal
lowerDiagonal = np.diag(array1, k = -1)

print('Array1:\n',array1)
print('Array1\'s main diagonal elements:\n', mainDiagonal)
print('Array1\'s diagonal elements above main diagonal:\n', upperDiagona )
print('Array1\'s diagonal elements below main diagonal:\n', lowerDiagonal)

输出

Array1:
[[1 2 3]
 [4 5 6]
 [7 8 9]]
Array1's main diagonal elements:
 [1 5 9]
Array1's diagonal elements above main diagonal:
 [2 6]
Array1's diagonal elements below main diagonal:
 [4 8]

相关方法

diagflat() - 创建一个二维数组，并将展平的输入作为其对角线。

import numpy as np

# create a 2D array
array1 = np.arange(1,5).reshape(2, 2)

# create diagonal array using diagflat()
mainDiagonal1 = np.diagflat(array1)

# create diagonal array using diag()
mainDiagonal2 = np.diag(array1.flatten())

print('Array1:\n',array1)
print('Array1\'s main diagonal elements:\n',mainDiagonal1)
print('Equivalent diag method:\n',mainDiagonal2)

输出

Array1:
 [[1 2]
 [3 4]]
Array1's main diagonal elements:
 [[1 0 0 0]
 [0 2 0 0]
 [0 0 3 0]
 [0 0 0 4]]
Equivalent diag method:
 [[1 0 0 0]
 [0 2 0 0]
 [0 0 3 0]
 [0 0 0 4]]

正如您所看到的，diagflat() 自动展平了 2D 数组，并创建了一个以展平数组元素作为其对角线的数组。

在 diag() 的情况下，我们手动使用了 flatten() 方法。