如果符合某些条件，则对来自不同数据帧的两个值求和

Question

如果满足某些条件，我想总结两列，每列在不同的框架中

数据帧1：

desk    Type   total_position
desk1   ES     786.0
desk1   ES1    100
desk2   ES1    0
desk2   ES2    10
desk3   ES     0
desk4   ES1    0
desk5   ES     -757

数据帧2：

desk    Type total_position
desk1   ES   -758.0
desk2   ES    0
desk3   ES   -29
desk4   ES    0.0
desk5   ES    786.0

如果第一个数据帧中的类型是“ES”，并且它是同一张桌子，我想对这两个位置求和

我该怎么做

预期答案

desk    Type   total_position
desk1   ES     29
desk2   ES1    0
desk3   ES     -29
desk4   ES1    0
desk5   ES     29

Answer 1

我会{}然后{}：

df1['total_position'] = (df1['total_position'].add(
               df1['desk'].map(df2.set_index('desk')['total_position']))
print(df1)

---

    desk Type  total_position
0  desk1   ES            28.0
1  desk2  ES1             0.0
2  desk3   ES           -29.0
3  desk4  ES1             0.0
4  desk5   ES            29.0

编辑类型：

m = (df1['desk'].map(df2.set_index('desk')['total_position'])
    .where(df1['Type'].eq('ES')).fillna(0))

df1['total_position'] = df1['total_position'].add(m)
print(df1)

    desk Type  total_position
0  desk1   ES            28.0
1  desk2  ES1             0.0
2  desk3   ES           -29.0
3  desk4  ES1             0.0
4  desk5   ES            29.0

Answer 2

试试这个：

通过索引添加，并使用combine_first更新df1中缺少的值

df1.set_index('desk').add(df2.set_index('desk')).combine_first(df1.set_index('desk'))

注意：这是基于一个天真的假设，即df1中的桌面与df2中的桌面相同

      Type  total_position
desk        
desk1   ES      28.0
desk2   ES1     0.0
desk3   ES     -29.0
desk4   ES1     0.0
desk5   ES      29.0

Answer 3

使用where

>>> df['total_position'] = (df.total_position+df1.total_position).where(df.desk.eq(df1.desk) & df.Type.eq('ES'), 0)
>>> df
    desk Type  total_position
0  desk1   ES            28.0
1  desk2  ES1             0.0
2  desk3   ES           -29.0
3  desk4  ES1             0.0
4  desk5   ES            29.0