how to crate the group by in pandas only in one level

问题

I am importing below df3 dataframe in my excel file and want to grouby only Name and rest dublicate data should reflect as below .

Note (Each Month data will be added as per month wise. )

Df3 =pd.read_Excel('Data')
print (df3)      

Name    ID  Month   Shift

Jon     1   Feb     A
Jon     1   Jan     B
Jon     1   Mar     C
Mike    1   Jan     A
Mike    1   Jan     B
Jon     1   Feb     C
Jon     1   Jan     A

and i want to have output like as below in the same formate . Please help me on same as im stuck here .

Will be greatfull for help and support .

回答1:

You can achieve it by

df=df.iloc[pd.to_datetime(df.Month,format='%b').argsort()]
df=pd.concat([pd.DataFrame({'Month':[x] }).append(y).fillna('').append(pd.DataFrame(dict.fromkeys(y.columns,['']))) for x , y in df.groupby('Name')]).drop('Name',1).iloc[:-1]

---

print(df)

 Month ID Shift
0   Jon         
1   Jan  1     B
6   Jan  1     A
0   Feb  1     A
5   Feb  1     C
2   Mar  1     C
0               
0  Mike         
3   Jan  1     A
4   Jan  1     B

回答2:

Code

#creating sample data as per requirement
import pandas as pd 
df = pd.DataFrame({'Name':['Jon','Jon','Jon','Mike','Mike','Jon','Jon'],'ID':[1,1,1,1,1,1,1], 'Month':['Feb','Jan','Mar','Jan','Jan','Feb','Jan'], 'Shift':['A','B','C','A','B','C','A']})
#display data
df

df['Month'] = pd.Categorical(df['Month'],categories=['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec'],ordered=True)
df = df.sort_values(['Name','Month']).reset_index(drop=True)
#display final data
df

I hope this would be helpful... : )

回答3:

Heres another solution using a list comp and df.duplicated with .loc for assignment.

import numpy as np
df = pd.read_excel(file,sheet_name=yoursheet)

#order the months. 

df['Month'] = pd.Categorical(df['Month'],
               pd.to_datetime(df['Month'],format='%b').drop_duplicates().sort_values().dt.strftime('%b'))



df = df.sort_values(['Month']).reset_index(drop=True)

df1 = pd.concat([data.append(data.iloc[0]) for name,data in df.groupby('Name')])

df1.loc[df1.duplicated(keep='last'),1:] = ''

df1['Name'] = np.where(df1['Month'].ne(''),df1['Month'],df1['Name'])

final = df1.drop('Month',1)

---

   Name ID Shift
0   Jon         
3   Jan  1     A
4   Feb  1     A
5   Feb  1     C
6   Mar  1     C
0   Jan  1     B
1  Mike         
2   Jan  1     B
1   Jan  1     A

来源：https://stackoverflow.com/questions/62612488/how-to-crate-the-group-by-in-pandas-only-in-one-level