R语言：将仅出现一次的Group观测重编码为Unaffiliated

先还原你的示例DataFrame：

import pandas as pd

df = pd.DataFrame({
    'Observation': [1, 2, 3, 4, 5, 6, 7],
    'X1': [2, 6, 8, 1, 2, 7, 2],
    'X2': [4, 3, 4, 3, 8, 5, 4],
    'Group': [1, 2, 2, 3, 4, 5, 5]
})

方法一：用transform快速匹配频次 #

通过transform('value_counts')给每一行匹配对应Group的出现次数，再用where方法替换符合条件的值：

# 给每行匹配对应Group的出现次数
group_counts = df['Group'].transform('value_counts')
# 仅保留出现次数>1的Group，其余替换为"Unaffiliated"
df['Group'] = df['Group'].where(group_counts > 1, 'Unaffiliated')

方法二：用map结合频次统计 #

先统计每个Group的出现次数，再通过lambda函数判断替换：

# 统计各Group的出现次数
counts = df['Group'].value_counts()
# 遍历每个Group值，判断后替换
df['Group'] = df['Group'].map(lambda x: x if counts[x] > 1 else 'Unaffiliated')

执行完任意一种方法后，输出的DataFrame就会和你期望的一致：

Observation  X1  X2          Group
0            1   2   4  Unaffiliated
1            2   6   3             2
2            3   8   4             2
3            4   1   3  Unaffiliated
4            5   2   8  Unaffiliated
5            6   7   5             5
6            7   2   4             5

内容的提问来源于stack exchange，提问作者flâneur