I have a numpy array of arrays, where each sub-array represents a group of elements. I want to compute the length of each sub-array and store the result as a new numpy array. Currently, I use a standard for loop, but I would like to vectorize this operation for better performance using only numpy. Example import numpy as np indices = np.array([np.array([1, 2, 3]), np.array([4, 5]), np.array([6, 7, 8, 9])], dtype=object) group_lengths = np.array([len(group) for group in indices]) print(group_lengths) Expected output [3 2 4] I am looking for a fully vectorized solution that avoids the explicit loop. How can this be achieved with numpy? Update Thank you all for your insights and clarifications! I now understand that fully vectorizing my use case isn't possible. I'll stick with my current solution, as it seems to be the most efficient approach given the circumstances. Continue reading...
