groupby.agg('sum).reset_index()妙用

@python
data
Out[22]: 
   data key1   key2
0     1    a    one
1     2    a    two
2     3    b    one
3     4    c  three
4     5    c    one

data.groupby(['key1','key2']).agg('sum').reset_index()
Out[29]: 
  key1   key2  data
0    a    one     1
1    a    two     2
2    b    one     3
3    c    one     5
4    c  three     4

使用groupby().agg().reset_index()可以使dataframe数据规整