如何将网格格式的数据转换为数组，反之亦然：Python

Question

我有一个数据集 data = f(x,y)。这个数据以网格格式可用。我想要将其转换为一个数组，使得每一行都按照 [x,y,data] 格式排序。另外，如何进行逆转换呢？

import numpy as np

x = np.arange(-5, 5, 0.1)
y = np.arange(-4, 4, 0.1)
xx, yy = np.meshgrid(x, y)
data = np.sin(xx**2 + yy**2) / (xx**2 + yy**2)

print(data)  #Now it is in meshgrid format

- Alam

Answer 1

您可以使用以下方法来重塑数据：

1. 使用网格格式

import numpy as np

x = np.arange(-5, 5, 2)
y = np.arange(-4, 4, 2)
xx, yy = np.meshgrid(x, y)
data = np.sin(xx**2 + yy**2) / (xx**2 + yy**2)

print(data)  # In meshgrid format

输出：

[[-0.00386885 -0.00529407 -0.05655279 -0.05655279 -0.00529407]
 [-0.02288393  0.03232054 -0.19178485 -0.19178485  0.03232054]
 [-0.00529407  0.04579094  0.84147098  0.84147098  0.04579094]
 [-0.02288393  0.03232054 -0.19178485 -0.19178485  0.03232054]]

2.以列格式：

z = len(x)*len(y)
a = np.reshape(xx, z)
b = np.reshape(yy, z)
c = np.reshape(data, z)

new = np.c_[a, b, c]

print(new) # In an array ordered like [x, y, data]

输出：

[[-5.00000000e+00 -4.00000000e+00 -3.86884558e-03]
 [-3.00000000e+00 -4.00000000e+00 -5.29407000e-03]
 [-1.00000000e+00 -4.00000000e+00 -5.65527936e-02]
 [ 1.00000000e+00 -4.00000000e+00 -5.65527936e-02]
 [ 3.00000000e+00 -4.00000000e+00 -5.29407000e-03]
 [-5.00000000e+00 -2.00000000e+00 -2.28839270e-02]
 [-3.00000000e+00 -2.00000000e+00  3.23205413e-02]
 [-1.00000000e+00 -2.00000000e+00 -1.91784855e-01]
 [ 1.00000000e+00 -2.00000000e+00 -1.91784855e-01]
 [ 3.00000000e+00 -2.00000000e+00  3.23205413e-02]
 [-5.00000000e+00  0.00000000e+00 -5.29407000e-03]
 [-3.00000000e+00  0.00000000e+00  4.57909428e-02]
 [-1.00000000e+00  0.00000000e+00  8.41470985e-01]
 [ 1.00000000e+00  0.00000000e+00  8.41470985e-01]
 [ 3.00000000e+00  0.00000000e+00  4.57909428e-02]
 [-5.00000000e+00  2.00000000e+00 -2.28839270e-02]
 [-3.00000000e+00  2.00000000e+00  3.23205413e-02]
 [-1.00000000e+00  2.00000000e+00 -1.91784855e-01]
 [ 1.00000000e+00  2.00000000e+00 -1.91784855e-01]
 [ 3.00000000e+00  2.00000000e+00  3.23205413e-02]]

- Nathaniel

Answer 2

您可以使用以下两个表达式：

# Convert from mesh to rows:
row_format = np.stack([z.ravel() for z in (xx, yy, data)], axis=1)

# Convert from rows to mesh:
xx, yy, data = np.swapaxes(row_format, 0, 1).reshape(3, len(y), len(x))

- a_guest