Using np.where with multiple conditions leads to incorrect indexing in NumPy 1.24
I'm working on a personal project and This might be a silly question, but I'm upgrading from an older version and I'm stuck on something that should probably be simple. I'm trying to use `np.where` to select elements from a NumPy array based on multiple conditions, but I'm running into unexpected results. My array contains the following data: ```python import numpy as np data = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]) ``` I want to select elements greater than 2 and less than 8. My initial approach was: ```python result = np.where((data > 2) & (data < 8)) ``` This returns the indices of the elements, which is fine, but when I try to use these indices to retrieve the actual values, I get an unexpected output: ```python selected_values = data[result] print(selected_values) ``` The output is: ``` array([3, 4, 5, 6, 7]) ``` While this is partially what I expected, I thought that `np.where` would directly give me the selected values instead of the indices. I checked if I could extract values directly, but it seems I am misusing the function. Additionally, I tried using `np.extract` as follows: ```python selected_values_extract = np.extract((data > 2) & (data < 8), data) ``` This gives me the output: ``` array([3, 4, 5, 6, 7]) ``` It seems to work, but it's slightly less intuitive than I expected and doesn’t maintain the original shape of the array. I want to retrieve these values while keeping the original shape, replacing the values that do not meet the condition with `np.nan`. Is there a more straightforward way to achieve this in NumPy 1.24, or am I misinterpreting how these functions should be used? Any guidance would be appreciated! My development environment is Ubuntu. I'm working with Python in a Docker container on Windows 10. Thanks, I really appreciate it! For reference, this is a production mobile app. Is there a simpler solution I'm overlooking?