row_missing_idx = np.flatnonzero(mask.any(axis=1))