np.concatenate yielding advanced patterns with masked arrays in NumPy
I'm wondering if anyone has experience with I need help solving I'm working on a project and hit a roadblock. I'm sure I'm missing something obvious here, but I'm working with masked arrays in NumPy and have encountered an scenario when attempting to concatenate them. I have two masked arrays that I want to combine using `np.concatenate`, but the result isn't what I expect. My masked arrays are structured like this: ```python import numpy as np # Creating two masked arrays mask1 = np.array([1, 2, 3]) mask2 = np.array([4, 5, 6]) masked_array1 = np.ma.masked_array(mask1, mask=[0, 1, 0]) # Masks the second element masked_array2 = np.ma.masked_array(mask2, mask=[1, 0, 0]) # Masks the first element ``` When I try to concatenate these arrays: ```python combined = np.concatenate((masked_array1, masked_array2)) ``` I expect the result to maintain the masked elements correctly, but instead, I get the following output: ```python masked_array(data=[1, --, 3, --, 5, 6], mask=[False, True, False, True, False, False]) ``` The masked values seem to have shifted positions, which is not what I anticipated. I tried using `np.ma.concatenate` as well, but it produced the same result. How can I concatenate these masked arrays without losing their masking structure or altering their intended layout? I'm currently using NumPy version 1.21.2. Any help or insights would be appreciated! Any ideas what could be causing this? Am I missing something obvious? Is there a better approach? I'm using Python 3.10 in this project. Has anyone dealt with something similar?