1>>> df["rank"] = df.groupby("group_ID")["value"].rank("dense", ascending=False)
2>>> df
3 group_ID item_ID value rank
40 0S00A1HZEy AB 10 2
51 0S00A1HZEy AY 4 3
62 0S00A1HZEy AC 35 1
73 0S03jpFRaS AY 90 1
84 0S03jpFRaS A5 3 5
95 0S03jpFRaS A3 10 2
106 0S03jpFRaS A2 8 4
117 0S03jpFRaS A4 9 3
128 0S03jpFRaS A6 2 6
139 0S03jpFRaS AX 0 7
14