python - 如何将 cumcount 应用于两列？

Question

我有一个包含名称和日期的数据框。我想创建一个计数列，它只会在日期不同时递增。请看下面第三栏:

Name    Date          COLUMN I NEED
----    ----          -------------
Bob     11-01-2019          1
Bob     11-01-2019          1
Bob     11-20-2019          2
Mike    12-01-2019          1
Mike    12-02-2019          2
Mike    12-03-2019          3
Steve   01-01-2019          1
Steve   01-01-2019          1

我尝试使用:

df['COLUMN RESULT'] = df.groupby(['Name'])['Date'].cumsum() + 1

Name    Date          COLUMN RESULT
----    ----          -------------
Bob     11-01-2019          1
Bob     11-01-2019          2
Bob     11-20-2019          3
Mike    12-01-2019          1
Mike    12-02-2019          2
Mike    12-03-2019          3
Steve   01-01-2019          1
Steve   01-01-2019          2

但是不管是什么日期它都会递增。感谢您的帮助，我很感激!

Answer 1

使用

df['result'] = df.Date.ne(df.Date.shift()).groupby(df.Name).cumsum().astype(int)

---

    Name        Date  result
0    Bob  11-01-2019       1
1    Bob  11-01-2019       1
2    Bob  11-20-2019       2
3   Mike  12-01-2019       1
4   Mike  12-02-2019       2
5   Mike  12-03-2019       3
6  Steve  01-01-2019       1
7  Steve  01-01-2019       1