Question

我有两个data_frame，如下所示：

df_name:
   Student_ID  Name        DOB
0           1  Raju 1993-02-02
1           2  Indu 1987-01-04
2           3  Laya 2000-06-24

df_marks:
    Student_ID Subject  Int1/40  Int2/40
0            1     Eng       10       35
1            1     Tam       30       38
2            1     Mat       20       30
3            1     Sci       15       20
4            2     Eng       35       25
5            2     Tam       25       15
6            2     Mat       22       30
7            2     Sci       29       23
8            3     Eng       18       17
9            3     Tam       19       16
10           3     Mat       27       26

任务是创建一个data_frame（下一个），如果需要df_marks['Int1/40']，我需要在其中添加df_marks['Int2/40']和df_name['Student_ID'] == df_marks['Student_ID']

   Student_id  Name        DOB  Tam/50
0           1  Raju 1993-02-02     NaN
1           2  Indu 1987-01-04     NaN  
2           3  Laya 2000-06-24     NaN

我尝试过

df_out['Tam/50'] = df_marks[['Int1/40','Int2/40']].sum(axis=1).where(df_marks['Subject']==df_out['Student_id'])

但是它给出的错误是

ValueError: Can only compare identically-labeled Series objects

我们有什么简单的方法吗？

关于，迪帕克·达什（Deepak Dash）

Answer 1

将DataFrame.join与汇总的sum一起用于df_name中的新列：

df_marks['Tam/50'] = df_marks[['Int1/40','Int2/40']].sum(axis=1)
df_name = df_name.join(df_marks.groupby('Student_ID')['Tam/50'].sum(), on='Student_ID')
print (df_name)
   Student_ID  Name         DOB  Tam/50
0           1  Raju  1993-02-02     198
1           2  Indu  1987-01-04     204
2           3  Laya  2000-06-24     123

或者没有帮助列的解决方案：

s = (df_marks[['Int1/40','Int2/40']].sum(axis=1)
                                    .groupby(df_marks['Student_ID'])
                                    .sum()
                                    .rename('Tam/50'))

df_name = df_name.join(s, on='Student_ID')
print (df_name)
   Student_ID  Name         DOB  Tam/50
0           1  Raju  1993-02-02     198
1           2  Indu  1987-01-04     204
2           3  Laya  2000-06-24     123

Answer 2

您可以使用pd.merge来匹配Student_ID上的两个数据帧。然后使用groupby和sum：

In [574]: res = pd.merge(df_name, df_marks,on='Student_ID')
In [592]: r = res.groupby(['Student_ID', 'Name', 'DOB'])[['Int1/40','Int2/40']].sum(1).reset_index()

In [594]: r['Tam/50'] = r['Int1/40'] + r['Int2/40']
In [604]: r.drop(['Int1/40', 'Int2/40'], 1, inplace=True)

In [605]: r
Out[605]: 
   Student_ID  Name         DOB  Tam/50
0           1  Raju  1993-02-02     198
1           2  Indu  1987-01-04     204
2           3  Laya  2000-06-24     123

Data_Frame错误-ValueError：只能比较标记相同的Series对象

2 个答案: