Hi, I would suggest to extract the final rank from ranking result (which is calculated by passing only the relevant attributes to the ranking algo) and append it to the original dataframe which contains all of the additional attributes which weren’t used for ranking.