Get ranking within Pandas.DataFrame with ties as possibility-CodePudding

I have the following example

import pandas as pd

names = ['a', 'b', 'c', 'd', 'e']
points = [10, 15, 15, 12, 20]

scores = pd.DataFrame({'name': names,
                       'points': points})

I want to create a new column called position that specifies the relative position of a player. The player with the most points is #1.

I sort the df using

scores = scores.sort_values(by='points', ascending=False)

If there is a tie (same number of points) I want position to be the T and the corresponding position. In my example the position of b and c is T2.

Desired output:

  name  points  position
     e      20         1
     b      15        T2
     c      15        T2
     d      12         3
     a      10         4

Thank you

CodePudding user response：

I would use pandas.Series.rank :

# is there a tie ?
m = scores["points"].duplicated(keep=False)

# calculate rankings
s = scores["points"].rank(method="dense", ascending=False)

scores["position"] = (
                        s.where(~m, s.astype(str).radd("T"))
                            .astype(str)
                            .replace(".0$", "", regex=True)
                     )

out = scores.sort_values(by="points", ascending=False)

# Output :

print(out)

  name  points position
4    e      20        1
1    b      15       T2
2    c      15       T2
3    d      12        3
0    a      10        4