How to make 2d complex number array in numpy?-CodePudding

I would like to make a 2d array of even distribution of complex numbers, a part of complex plane, for example (-1, 1i), (-1, -1i), (1, 1i), (1, -1i) with 20 numbers in each dimension.

I know I can do this for complex numbers in 1 d with np.linspace like this:

import numpy as np

complex_array = np.linspace(0, complex(1, 1), num = 11)
print(complex_array)

[0.  0.j, 0.1 0.1j, 0.2 0.2j, 0.3 0.3j, 0.4 0.4j,
 0.5 0.5j, 0.6 0.6j, 0.7 0.7j, 0.8 0.8j, 0.9 0.9j, 1.  1.j ]

But I can't get my head around how to produce this in two dimensions to get a part of a complex plane?

Some somewhat similar questions mention np.mgrid, but the examples are with reals and I would like the array to contain dtype=complex so my math keeps simple.

Maybe I am just missing something, and perhaps just a simple example would explain a lot..

CodePudding user response：

There is no magic about complex numbers - they are simply a way to express a two dimensional space. You could use np.meshgrid (see here) to define a two dimensional Cartesian grid and then combine the coordinates into complex numbers.

Create vectors which will span the two dimensional grid (or complex plane)

real_points = np.linspace(0,1,num=11) imag_points = np.linspace(0,1,num=11)

Create 2-D coordinate arrays

real_grid, imag_grid = np.meshgrid(real_points, imag_points)

Combine into complex array:

complex_array = real_grid imag_grid * 1j

This produces a 11x11 complex128 array.

CodePudding user response：

You can use broadcasting to do that. For example:

result = np.linspace(0, 1j, num = 11).reshape(-1, 1)   np.linspace(0, 1, num = 11)

Using meshgrid also works but it is likely slower:

a, b = np.meshgrid(np.linspace(0, 1, num = 11), np.linspace(0, 1j, num = 11))
result = a   b