zl程序教程

您现在的位置是:首页 >  Javascript

当前栏目

Hive综合案例练习(中级)第二题:查询至少连续三天下单的用户

2023-04-18 14:25:00 时间

查询至少连续三天下单的用户

题目需求

查询订单信息表(order_info)中最少连续3天下单的用户id,期望结果如下:

user_id
101

题目解析

  • 第一步:用户一天可能存在多次下单的情况,所以需要对用户进行去重

去重方法如下:

--方法一:使用distinct
select
    distinct
    user_id,
    create_date
from order_info;

--方法二:使用group by
select
    user_id,
    create_date
from order_info
group by user_id, create_date;

--方法二:使用窗口函數
select
    user_id,
    create_date
from
(
    select
        user_id,
        create_date,
        row_number() over (partition by user_id,create_date) rn
    from order_info
)t1
where rn=1;
  • 第二步:具体见各个方案。

代码实现

  • 方案一

    select
        distinct user_id
    from
    (
        select
            user_id
        from
        (
            select
                user_id,
                create_date,
                date_sub(create_date, row_number() over (partition by user_id order by create_date)) diff
            from
            (
                select
                    user_id,
                    create_date
                from order_info
                group by user_id, create_date
            )t1
        )t2
        group by user_id
        having count(diff) >= 3
    )t3
    
  • 方案二

    select
        distinct user_id
    from
    (
        select
            user_id,
            datediff(lead2, create_date) diff
        from
        (
            select
                user_id,
                create_date,
                lead(create_date, 2, '9999-12-31') over (partition by user_id order by create_date) lead2
            from
            (
                select
                    user_id,
                    create_date
                from order_info
                group by user_id, create_date
            )t1
        )t2
    )t3
    where diff=2;
    
  • 方案三

    select
        distinct user_id
    from
    (
        select
            user_id,
            ts,
            count(*) over (partition by user_id order by ts range between 86400 preceding and 86400 following) cnt
        from
        (
            select
                user_id,
                unix_timestamp(create_date, 'yyyy-MM-dd') ts
            from
            (
                select
                    user_id,
                    create_date
                from order_info
                group by user_id, create_date
            )t1
        )t2
    )t3
    where cnt=3;