面试遇到的订单表sql的解决方案
Posted hongma
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了面试遇到的订单表sql的解决方案相关的知识,希望对你有一定的参考价值。
对于以下需求:用户表:users (user_id int)
订单表:order_tb(user_id int, or_time date, or_money double)
求以下用户:
一月下过单,二月份没有下过单的三月份下单用户
的如下指标:
三月份订单金额大于100的订单数,三月份第一笔订单和最后一笔订单的订单金额。(每个用户的哦)
网上求得如下解答,比自己之前写的几百行好多了。不知道有没有其它解决方案
with us1 as (select user_id,or_time,or_money,
sum(nvl(case when to_char(or_time,‘MM‘) = ‘01‘ then 1 end,0)) over (partition by user_id) m1,
sum(nvl(case when to_char(or_time,‘MM‘) = ‘02‘ then 1 end,0)) over (partition by user_id) m2,
sum(nvl(case when to_char(or_time,‘MM‘) = ‘03‘ then 1 end,0)) over (partition by user_id) m3,
sum(case when to_char(or_time,‘MM‘) = 03 and or_money > 100 then 1 end) over (partition by user_id) cnt
from order_tb
where or_time between to_date(‘20170101‘,‘YYYYMMDD‘) and to_date(‘20170331‘,‘YYYYMMDD‘)),
mo1 as (select distinct user_id,
case when to_char(or_time,‘MM‘) = 03 then first_value(or_money) over (partition by user_id order by or_time) end f1v,
case when to_char(or_time,‘MM‘) = 03 then last_value(or_money) over (partition by user_id order by or_time range between unbounded preceding and unbounded following) end f3l,
cnt from (select * from us1 where m2=0 and m1 >0 and m3 >0 and or_time between to_date(‘20170301‘,‘YYYYMMDD‘) and to_date(‘20170331‘,‘YYYYMMDD‘)))
以上是关于面试遇到的订单表sql的解决方案的主要内容,如果未能解决你的问题,请参考以下文章