通过组合其他列的信息来创建新列

Question

如何创建包含其他列信息的“消息”列？数据框已经以各种方式排序。

df   head        members   
0   Abba As      Ally       
1   Abba As      Apo
2   Abba As      Abba
3   Bella Bi     Bella
4   Bella Bi     Boo
5   Bella Bi     Brian
6   Abba As      Arra       
7   Abba As      Alya
8   Abba As      Abba

预期产出

df  head                           message
0   Abba As          Hi Abba, we invite you, Ally and Apo. Please use "Abba As" when arriving.
1   Bella Bi         Hi Bella, we invite you, Boo and Brian. Please use "Bella Bi" when arriving.
2   Abba As          Hi Abba, we invite you, Arra and Alya. Please use "Abba As" when arriving.

我尝试创建名字列：

df["head_first_name"] = df.head.str.split(" ").str[0]
df.loc[df.head_first_name.isin(df.members)

Answer 1

您可以将

groupby.apply

与自定义函数一起使用：

def message(g):
    head = g.name[0].split()[0]
    others = ' and '.join([m for m in g['members'] if m != head])
    return f'Hi {head}, we invite you, {others}. Please use "{g.name[0]}" when arriving.'

group = df['head'].ne(df['head'].shift()).cumsum()
out = (df.groupby(['head', group], sort=False)
         .apply(message)
         .droplevel(1)
         .reset_index(name='message')
      )

输出：

       head                                                                       message
0   Abba As     Hi Abba, we invite you, Ally and Apo. Please use "Abba As" when arriving.
1  Bella Bi  Hi Bella, we invite you, Boo and Brian. Please use "Bella Bi" when arriving.
2   Abba As    Hi Abba, we invite you, Arra and Alya. Please use "Abba As" when arriving.