Pergunta de entrevista da empresa Enterra Solutions

Given Pandas and NumPy, why would you sometimes prefer one over the other? How would you handle missing values in a dataset?