我有下表
| userid | eventid | urlclick |
|---|---|---|
| A | EID-1 | 0,0,0,0,0 |
| A | EID-1 | 0,1,2,3,0 |
| A | EID-1 | 0,1,0,1,0 |
| A | EID-2 | 1,0,0,0,0 |
| A | EID-2 | 0,0,2,4,1 |
| A | EID-2 | 0,0,0,1,0 |
我能够使用此代码将 urlclick 列转换为 np.array
time.urlclick = time['urlclick'].apply(lambda x: np.array(eval(x)))
现在我想按事件 id 对表格进行分组，并将 urlclikc 与axis = 0 相加
time.groupby(['userid','eventid']).urlclick.apply(np.array).sum(axis = 0)
所以结果应该是
EID-1 - 0,2,2,4,0
EID-2 - 1,0,2,5,1
答案 0 :(得分：0)
time.groupby(['userid','eventid']).urlclick.apply(lambda x: x.sum(axis = 0))