python - 根据 Pandas 中的条件连接多个值将单个列呈现为新列

Question

首先，我应该说我已经针对此类问题尝试过可能的解决方案，但找不到最接近的解决方案。我的数据框如下所示

Type    Run    Status    Message
P1      R1     OK 
P1      R2     NOK       Unable to connect
P1      R3     OK
P1      R4     NOK       Unable to fetch
P2      R1     OK
P2      R2     OK
P2      R3     NOK       Entry not present
P2      R4     NOK       Entry not present

我想将每个“Type”== NOK 的所有唯一错误消息连接到新列

Type    Run    Status    Message                Error
P1      R1     OK 
P1      R2     NOK       Unable to connect      Unable to connect, Unable to fetch
P1      R3     OK
P1      R4     NOK       Unable to fetch
P2      R1     OK
P2      R2     OK
P2      R3     NOK       Entry not present      Entry not present
P2      R4     NOK       Entry not present

任何线索都会有帮助

Answer 1

您可以使用groupby操作:

# identify NOK rows
m = df['Status'].eq('NOK')
# get index of first NOK per group
idx = m[m].groupby(df['Type']).cumcount().loc[lambda x: x==0].index

# initialize Error to empty string (optional)
df['Error'] = ''

# concatenate unique error messages per group
# and assign to first NOK per group
df.loc[idx, 'Error'] = (df.groupby('Type')['Message']
                          .agg(lambda g: ', '.join(dict.fromkeys(g.replace('', pd.NA).dropna())))
                          .tolist()
                       )

输出:

  Type Run Status            Message                               Error
0   P1  R1     OK               None                                    
1   P1  R2    NOK  Unable to connect  Unable to connect, Unable to fetch
2   P1  R3     OK               None                                    
3   P1  R4    NOK    Unable to fetch                                    
4   P2  R1     OK               None                                    
5   P2  R2     OK               None                                    
6   P2  R3    NOK  Entry not present                   Entry not present
7   P2  R4    NOK  Entry not present