问使用python在dataframe中按类别汇总过去12个月的数据

EN

import pandas as pd
import datetime

df1.Start_Date = pd.DatetimeIndex(df1.Start_Date)

today = datetime.date.today()
currentfirst = today.replace(day=1)
thirdMonth = currentfirst - pd.offsets.MonthBegin(3)
secondMonth = currentfirst - pd.offsets.MonthBegin(2)
firstMonth = currentfirst - pd.offsets.MonthBegin(1)

fst_label = firstMonth.strftime('%B')
snd_label = secondMonth.strftime('%B')
thd_label = thirdMonth.strftime('%B')

def monthly_vol(df, label, start_date, end_date):
    """Slices df1 into previous months and sums the volume of each change class."""
    if start_date is not None:
        df = df1[df1.Start_Date >= start_date]
    if end_date is not None:
        df = df[df.Start_Date < end_date]
    df_count = df.groupby('Change Class').count().drop(['Start_Date', 'Risk Level', 'Change Coordinator', 'Change Coordinator Group'], axis=1)
    return df_count

fst_month = monthly_vol(df1, fst_label, firstMonth, currentfirst)
snd_month = monthly_vol(df1, snd_label, secondMonth, firstMonth)
thd_month = monthly_vol(df1, thd_label, thirdMonth, secondMonth)

def month_merge(df1, df2, df3):
    """Merges monthly dataframes together."""
    new_df = pd.merge(df1, df2, left_index=True, right_index=True).merge(df3, left_index=True, right_index=True)
    new_df.columns = [fst_label, snd_label, thd_label]
    print(new_df)
    return new_df

monthly_vol = month_merge(fst_month, snd_month, thd_month)

              May  April  March
Change Class                   
Emergency      36     36     32
Expedited      17     24     35
Normal        182    146    134
Standard      256    210    267

              May  MayAVG  April  AprilAVG   March  MarchAVG
Change Class                   
Emergency      36   7.33   36     8.65       32     6.84
Expedited      17   3.46   24     5.77       35     7.48
Normal        182   37.07  146    35.10      134    28.63
Standard      256   52.14  10     50.48      267    57.05