pandas 语句

1、对于时间格式数据的处理

有些时候time_stamp是object格式，提取相应的日期，小时，星期等：

方法1

 from datetime import datetime

 user_trad['time_stamp']=user_trad['time_stamp'].apply(lambda x: datetime.strptime(x,'%Y-%m-%d %H:%M'))
user_trad['hour']=user_trad['time_stamp'].dt.hour
user_trad['date']=user_trad['time_stamp'].dt.date
user_trad['weekday']=user_trad['time_stamp'].dt.weekday 

shop_hour_user_cnt['hour_cate']=shop_hour_user_cnt['hour'].apply(am_pm_hour)

 

方法2

 #添加小时
   join_table['hour']=int(str(join_table['time_stamp'][0])[11:13])
   join_table['hour_type']=join_table['hour'].map(am_pm_hour)
   #加上每天是星期几
   join_table['date']=(join_table['time_stamp'][0])[0:4]+str(join_table['time_stamp'][0])[5:7]+str(join_table['time_stamp'][0])[8:10]
   join_table['weekday']=datetime.strptime(str(join_table['date'][0]),"%Y%m%d").weekday() +1

 

 

2、

# 按每个小时去重到店客户
tmp=user_trad.drop_duplicates(['shop_id','user_id','hour'])