在新列中插入组的总和

Question

我有一个简单的DataFrame：

gov dis value1  value2
a   a_1 8   8
a   a_2 7   18
a   a_3 3   2
a   a_4 12  12
b   b_1 4   11
b   b_2 16  9
b   b_3 17  12
b   b_4 7   9
b   b_5 15  11
b   b_6 13  16

我可以轻松地创建value1中所有值的总和。你知道吗

per_t['sum_v1'] = sum(per_t['value1'])

生产：

    gov dis value1  value2  sum_v1
0   a   a_1 8   8   102
1   a   a_2 7   18  102
2   a   a_3 3   2   102
3   a   a_4 12  12  102
4   b   b_1 4   11  102
5   b   b_2 16  9   102
6   b   b_3 17  12  102
7   b   b_4 7   9   102
8   b   b_5 15  11  102
9   b   b_6 13  16  102

是否可以将其更改为计算每个saygov的总和？我尝试了groupby和agg函数，但没有成功。如果我尝试：

per_t['sum_gov'] = per_t.groupby('gov')['value1'].sum()

表格内容如下：

gov dis value1  value2  sum_v1  sum_gov
0   a   a_1 8   8   102 NaN
1   a   a_2 7   18  102 NaN
2   a   a_3 3   2   102 NaN
3   a   a_4 12  12  102 NaN
4   b   b_1 4   11  102 NaN
5   b   b_2 16  9   102 NaN
6   b   b_3 17  12  102 NaN
7   b   b_4 7   9   102 NaN
8   b   b_5 15  11  102 NaN
9   b   b_6 13  16  102 NaN

Answer 1

你可以的

df['sum_gov'] = df.groupby('gov')['value1'].transform(sum)

见docs：

The transform method returns an object that is indexed the same (same size) as the one being grouped. Thus, the passed transform function should return a result that is the same size as the group chunk.