python-3.x - Pandas 组合字符串行

Question

我有一个像这样的 Pandas 数据帧:

Index   User    Message
1       User1   This is
2       User1   a test.
3       User2   The test
4       User2   is received.
5       User1   This is a complete sentence.
6       User2   This is another complete sentence.
7       User1   Finished.

我需要一种方法来 当用户发送多条消息时合并消息 在另一个之前，同时保持原来的顺序。
期望的输出:

Index   User    Message
1       User1   This is a test.
2       User2   The test is received
3       User1   This is a complete sentence.
4       User2   This is another complete sentence.
5       User1   Finished.

Pandas 可以这样做吗？

Answer 1

尝试:

df = (df.groupby([df['User'].ne(df['User'].shift(1)).cumsum().values, 'User'])['Message']
        .agg(' '.join).reset_index(level=1))