使用子字符串筛选listlike格式的dataframe列

Question

我有以下人员编号数据框（唯一且不重复）以及这些人员以类似列表的格式压缩的角色（针对具有多个角色的人员），如下所示：

         Person Number                                          Role Name
0          10000170             [HR Business Partner, HR Administrator, HR Administrator Limited]
1          10000479             [HR Administrator, HR Administrator Reporting]
2          10001347             [HR Business Partner, HR Administrator]
3          10003004             [HR Business Partner Approver, HR Business Partner]
4          10003048             [HR Administrator, HR Administrator Reporting]
5          10062441             [HR Business Partner Security Limited, HR Administrator Limited]
6          10062862             [HR Business Partner, HR Business Partner Approver]
7          10063101             HR Business Partner
8          10063833             HR Administrator

我希望按以下方式筛选此数据帧：

1-返回“角色名称”列采用类似列表格式的所有行：

输出：

         Person Number                                          Role Name
0          10000170             [HR Business Partner, HR Administrator, HR Administrator Limited]
1          10000479             [HR Administrator, HR Administrator Reporting]
2          10001347             [HR Business Partner, HR Administrator]
3          10003004             [HR Business Partner Approver, HR Business Partner]
4          10003048             [HR Administrator, HR Administrator Reporting]
5          10062441             [HR Business Partner Security Limited, HR Administrator Limited]
6          10062862             [HR Business Partner, HR Business Partner Approver]

2-以返回所有员工行的方式筛选此数据框，其中“角色名称”值中可能包含单词“业务合作伙伴”和单词“管理员”，返回以下内容：

输出：

         Person Number                                          Role Name
0          10000170             [HR Business Partner, HR Administrator, HR Administrator Limited]
2          10001347             [HR Business Partner, HR Administrator]
5          10062441             [HR Business Partner Security Limited, HR Administrator Limited]

我如何实现每个输出

Answer 1

一种可能的想法是使用自定义函数从列Role Name生成一个布尔掩码，方法是使用^{}，最后使用此掩码过滤数据帧中的行：

def fx(v):
    if not isinstance(v, list):
        return False

    roles = ','.join(v).lower()
    return 'business partner' in roles and 'administrator' in roles

df1 = df[df['Role Name'].apply(fx)]

结果:

# print(df1)
  Person Number                                          Role Name
0      10000170  [HR Business Partner,  HR Administrator,  HR A...
2      10001347           [HR Business Partner,  HR Administrator]
5      10062441  [HR Business Partner Security Limited,  HR Adm...