1. #!/usr/bin/env python
    2. # coding: utf-8
    5. import numpy as np
    6. import pandas as pd
    10. df  = pd.read_csv(r"D:\Mywork\DB\MANIFEST_CONSUME_RECORD_202206081957.csv",
    11.                   usecols =['CREATE_TIME','COMPANY_NAME'])
    12. df
    16. #转换为时间类型
    17. df['CREATE_TIME'] = pd.to_datetime(df['CREATE_TIME'])
    22. #把时间格式改为年月日不要时分秒
    23. df['yearmonthday'] = df['CREATE_TIME'].apply(lambda x: x.strftime('%Y-%m-%d'))
    27. #去掉两个字段的重复项
    28. df1 = df.drop_duplicates(subset=['COMPANY_NAME','yearmonthday'],keep='first')
    32. #按两个字段排序
    33. df2 = df1.sort_values(['COMPANY_NAME','yearmonthday'])
    37. #生成新的一列
    38. df2['newtime'] = df2.groupby('COMPANY_NAME')['yearmonthday'].shift(-1)
    39. df2
    43. #替换nan值
    44. df3 = df2.fillna('2022-06-01')
    45. df3
    49. #转换时间
    50. df3['newtime2']=pd.to_datetime(df3['yearmonthday'])
    51. df3['newtime3']=pd.to_datetime(df3['newtime'])
    52. df3
    56. df3['diff'] = df3['newtime3'] - df3['newtime2']
    57. df3
    61. group = df3.groupby('COMPANY_NAME').max()
    62. group
    65. group.to_csv("D:\Mywork\DB\diff3",sep=',')