The packet classification operation pandas

df = pd.read_excel('D:/data/data2/20190602.xlsx')

The following example operates on a single packet after the packet into a packet after the merger

df5 = pd.DataFrame()

for so, i in df.groupby(by='singer'):

    df3 = i.groupby(['Title', 'Artist']).agg({'Title': max）

    temp = df3

    df5 = df5.append(temp)

df5.to_excel('D:/data/0602.xlsx', index=False)

After nothing more than create an empty DataFrame in front of the classification processed in Add to that dataframe

The question I think for a long time