python - 使用pivot_table时同时对aggfunc应用两个numpy函数

Question

我有这张表:

user_id | datetime   | type
1       | 2015-01-01 | 1
1       | 2015-01-01 | 2
1       | 2015-01-01 | 2
1       | 2015-01-02 | 2
2       | 2015-01-01 | 2
2       | 2015-01-02 | 1
2       | 2015-01-02 | 2

我有这个pivot_table代码:

df = df.pivot_table('type', ['user_id'], ['datetime'], aggfunc=np.mean)

但是，我想同时应用两个 unique().sum() 函数来满足此条件，而不是 np.mean:

If there are both 1 and 2 during specific days per user, then I want to put 3, if there is only 1 for a specific day I want to put 2, etc.

例如，所需的输出如下:

user_id | 2015-01-01 | 2015-01-02 
1       | 3          | 2 
2       | 2          | 3 

有什么想法吗？

Answer 1

这是你想要的吗？

In [50]: df.pivot_table('type', ['user_id'], ['datetime'], aggfunc=lambda x: x.unique().sum())
Out[50]:
datetime  2015-01-01  2015-01-02
user_id
1                  3           2
2                  2           3