Unexpected NaN values when using np.polyfit with masked arrays in NumPy 1.25
I'm confused about I'm sure I'm missing something obvious here, but I'm working with an scenario when using `np.polyfit` with masked arrays in NumPy 1.25... I have a dataset where some of the x and y values are invalid and thus masked. When I try to fit a polynomial to this data, I end up with NaN values in the coefficients. Here's a snippet of my code: ```python import numpy as np import numpy.ma as ma # Sample data with some invalid entries x = np.array([1, 2, 3, 4, 5]) y = np.array([2.5, 3.5, np.nan, 5.5, 6.5]) # Creating a masked array for y, masking the NaN values masked_y = ma.masked_array(y, np.isnan(y)) # Trying to fit a polynomial of degree 1 coefficients = np.polyfit(x, masked_y, 1) print(coefficients) ``` When I run this code, I get the output: ``` [nan nan] ``` I've tried using `ma.compressed()` on `masked_y` prior to fitting, but that results in a ValueError due to shape mismatch. The documentation indicates that `np.polyfit` should handle masked arrays, but it seems to be failing in this instance. Is there a proper way to handle this or a workaround to avoid getting NaN coefficients? Any insights would be greatly appreciated! Any help would be greatly appreciated! This is part of a larger CLI tool I'm building. Is there a better approach? I'd love to hear your thoughts on this.