I have a dataframe with anywhere from 2-5 different fruits, and a single column for the count. There will never be repeats of fruits on each row in the dataframe. However, the combination of fruits could have duplicates. The count will always stay the same for these duplicates since it's giving the count of the combination of all the fruits. I want to keep just a unique row of the combination of fruits, and I want to remove the other duplicate combination rows. Of the duplicate combinations, it does not matter which row is retained and which is removed. If it makes it easier, we can sort by ascending order starting with first_fruit, second_fruit, and then third_fruit. Lastly, I only want to keep the top X rows in terms of count. The example below has 3 fruits. Keep only unique combinations. Row 2 would be removed since we sort by first_fruit, second_fruit, and then third_fruit in ascending order. Keep the top 2 rows in terms of count in descending order. Row 4 would be removed since it's 3rd in terms of count. Current Dataframe df = pd.DataFrame({'first_fruit': ['apple','orange','banana','cherry'], 'second_fruit': ['orange','apple','cherry','apple'], 'third_fruit': ['cherry','cherry','apple','orange'], 'count':['10','10','5','3']}) Desired Dataframe Output desired_output = pd.DataFrame({'first_fruit': ['apple','banana'], 'second_fruit': ['orange','cherry'], 'thid_fruit': ['cherry','apple'], 'count':['10','5']}) Please let me know if any clarification is needed. I appreciate the support! Continue reading...
