sql面试题20200607

表结构

以下是数据库中 4 个表的表结构：

表一：订单表（表名：table_order），主键是order_id+goods_id，用于记录每个订单的商品明细。同一个订单可能有多个商品，每条信息记录每个订单里不同商品的件单价和销售件数等，销售额=件单价price*销售件amount

备注：order_id：订单号，add_time：下单时间，user_id：客户号，goods_id：商品编号，price：件单价，amount：销售件数，department_id：品类编号

表二：品类表（表名：table_department），主键是department_id。

表三：客户表（表名：table_user），主键是user_id。

备注：user_id：客户号，user_name：客户姓名

表：购物车表（表名：table_cart），主键是user_id+goods_id+add_time，用于记录客户不同时点放入购物车的商品情况（商品ID、商品数量、商品删除时间、商品购买时间）

备注：客户主动删除商品或购买商品，时间分别记录在delete_time和buy_time两个字段，否则记录为空

需求

请在25 分钟内完成下面4 道试题：
1：请写出“计算2014/03/22-2014/04/30 每天的购买客户数、订单量、销售件数、销售额”的
SQL 代码，结果格式要求如下：

2：请写出“计算2014 年4 月各品类的销售额、晚上20-24 点销售额”的SQL 代码，结果格式
要求如下：

3：请写出“提取2014 年3-5 月销售额排名前三的客户信息（排名/客户号/客户姓名/总销售额
/购买品类）”的SQL 代码，结果格式要求如下：

4：特卖模式下，客户把商品放入购物车后，需在 20 分钟内完成购买，每加入一件新的商品，购物时长（即 20 分钟）将重新计算，若在 20 分钟内无放入购物车行为，购物车内商品会由系统自动释放。请结合购物车表的数据结构，写出计算 2014 年 3 月每日平均每件商品 hold 货时长（即商品的锁定时间）的逻辑和 SQL 代码，结果格式要求如下：

建表导数

$ vi create_table.sql

编辑保存如下内容：

--订单表
drop table if exists table_order;
create table table_order(
  order_id  string 
, user_id string 
, goods_id string
, add_time timestamp 
, price decimal(32,2)
, amount int
, department_id string
) row format delimited fields terminated by '\t';
insert into table_order values ('1236','111','20340','2014-03-22 01:19:00',398.00,1,'1');
insert into table_order values ('1236','111','12432','2014-03-22 01:19:00',18.00 ,2,'2');
insert into table_order values ('1648','333','56235','2014-04-11 21:00:00',399.00,3,'3');
insert into table_order values ('1648','333','42521','2014-04-11 21:00:00',290.00,1,'4');
insert into table_order values ('1648','333','24466','2014-04-11 21:00:00',279.00,2,'1');
insert into table_order values ('2060','222','25948','2014-03-30 13:21:00',190.00,2,'2');
insert into table_order values ('2472','222','20942','2014-04-22 09:59:00',380.00,1,'3');
insert into table_order values ('2472','222','98303','2014-04-22 09:59:00',299.00,1,'4');
insert into table_order values ('2472','222','39555','2014-04-22 09:59:00',848.00,2,'4');
insert into table_order values ('2884','444','20249','2014-05-07 17:53:00',599.00,1,'2');
insert into table_order values ('2884','444','39523','2014-05-07 17:53:00',699.00,1,'2');

--品类表
drop table if exists table_department;
create table table_department(
  department_id   string 
, department_name string 
) ;
insert into table_department values ('1','女装');
insert into table_department values ('2','女鞋');
insert into table_department values ('3','男装');
insert into table_department values ('4','男鞋');

--客户表
drop table if exists table_user;
create table table_user (
  user_id string
, user_name string
);
insert into table_user values('111','张敏');
insert into table_user values('222','黎佳');
insert into table_user values('333','杨峰');
insert into table_user values('444','催楠');

--购物车表
drop table if exists table_cart;
create table table_cart(
  user_id string
, goods_id string
, amount int
, add_time timestamp
, delete_time timestamp
, buy_time timestamp
);
insert into table table_cart values ('111','56235',2,'2014-10-22 10:01:00',null,'2014-10-22 10:20:00');
insert into table table_cart values ('111','24466',1,'2014-10-22 10:21:00','2014-10-22 10:29:00',null);	
insert into table table_cart values ('111','25948',3,'2014-10-22 10:24:00',null,null);
insert into table table_cart values ('111','39555',1,'2014-10-22 21:20:00','2014-10-22 21:26:00',null);	
insert into table table_cart values ('111','98303',4,'2014-10-22 21:24:00',null,null);
insert into table table_cart values ('111','20249',2,'2014-10-22 21:39:00',null,null);

$ hive -f create_table.sql

执行建表导数脚本

查询脚本

需求1

–1.计算 2014/03/22-2014/04/30 每天的购买客户数、订单量、销售件数、销售额

select 
  count(distinct user_id) as user_id_num
 ,count(distinct order_id) as order_id_num 
 ,sum(amount) as amount_num
 ,sum(price*amount) as total_price
from 
 table_order
where date(add_time) 
between '2014-03-22' and '2014-04-30';

查询结果

需求2

–2.计算 2014 年 4 月各品类的销售额、晚上 20-24 点销售额

select 
   b.department_name as department_name
 , sum(a.price*a.amount) as total_sale
 , sum(if(hour(a.add_time) between 20 and 24,a.price*a.amount,0)) as evening_total_sale
from 
 table_order a
join table_department b
on a.department_id = b.department_id
group by b.department_name;

查询结果

需求3

–提取 2014 年 3-5 月销售额排名前三的客户信息（排名/客户号/客户姓名/总销售额/购买品类）

select
    t1.rn as rn
  , t1.user_id  as user_id
  , t2.user_name as user_name
  , t1.total_sale as total_sale
  , t1.department_name_col as department_name_col
from 
  (select 
        a.user_id
      , sum(price*amount) as total_sale
      , concat_ws('、',collect_set(b.department_name)) AS department_name_col
      , row_number() over(order by sum(price*amount) desc) as rn
  from 
       table_order a
  left join table_department b
  on a.department_id = b.department_id
  group by 
     a.user_id
  ) t1 --通过客户id将品类数据行转列
left join table_user t2 --取客户名称
on t1.user_id = t2.user_id
where t1.rn <= 3;

查询结果

需求4

–计算 2014 年 3 月每日平均每件商品 hold 货时长（即商品的锁定时间）

select
    to_date(add_time) as dt
  , sum(duration_time)/(count(1)*60.00) as per_duration_min --平均时长(分钟)
from
  (select
      add_time
    , unix_timestamp(coalesce(delete_time,buy_time,each_end_time)) - unix_timestamp(add_time) as duration_time -- 每条的时长
   from
       (select
          user_id
        , add_time
        , delete_time
        , buy_time
        , min(cast(end_time as timestamp)) over(partition by user_id order by add_time desc rows between unbounded preceding and current row) as each_end_time --取每批的自动清空时间
       from
           (
           select 
                 user_id
               , add_time
               , delete_time
               , buy_time
               , case when unix_timestamp(next_time) - unix_timestamp(add_time) > 20*60 or next_time is null then from_unixtime(unix_timestamp(add_time)+ 20*60) end as end_time --只对分组最后一条打上结束时间
           from
               (select
                   user_id
                 , add_time
                 , delete_time
                 , buy_time
                 , lead(add_time) over(partition by user_id order by add_time) as next_time --下一条商品加入时间
               from table_cart
               ) t1
           )t2
       )t3
  )t4
group by to_date(add_time);

查询结果