hive业务场景

表一：orders（订单表）

create table orders(
order_id string,
user_id string,
eval_set string,
order_number string,
order_dow string,
order_hour_of_day string,
days_since_prior_order string
)
row format delimited fields terminated by ',' lines terminated by '\n';

 

表二：prior（商品表）

create table priors(
order_id string,
product_id string,
add_to_cart_order string,
reordered string
)
row format delimited fields terminated by ',' lines terminated by '\n';

业务1：

查询每个用户有多少个订单？

select user_id,count(1) as order_cnt 
from orders 
group by user_id 
order by order_cnt desc 
limit 10;

业务2：

每个用户平均每个订单平均是多少商品 ？

思路，某个用户平均每个订单平均的商品数量=某个用户一共购买的商品数量/一共购买的订单数量

1.先在商品表算出一个订单有多少个商品

select order_id,count(1) as pri_cnt
from priors 
group by order_id

2.将商品表的order_id和订单表的order_id关联起来，把订单中的产品数量信息带到用户信息里

select odr.user_id,pri.order_id,pri_cnt
from orders odr
join
(
select order_id,count(1) as pri_cnt
from priors 
group by order_id) pri
on odr.order_id=pri.order_id
limit 10;

3.求和，算出每个用户总共购买多少商品

select user_id,sum(pri_cnt)
from(
select odr.user_id,pri.order_id,pri_cnt
from orders odr
join
(
select order_id,count(1) as pri_cnt
from priors 
group by order_id) pri
on odr.order_id=pri.order_id
)t
group by user_id
limit 10;

4.求和，算出某个用户一共购买多少订单

每个用户有多少个订单：
  select user_id,count(1)
  from orders 
  group by user_id;

5.求平均

  select user_id,
  sum(prod_cnt)/count(1) as sc_prod,
  avg(prod_cnt) as avg_prod 
  from (select * from orders where eval_set='prior')od --如果不是prior统计为0
  join (
  select order_id,count(1) as prod_cnt 
  from priors 
  group by order_id
  limit 10000)pro
  on od.order_id=pro.order_id
  group by user_id
  limit 10;

业务3：每个用户在一周中的购买订单的分布(列转行)

user_id,dow0,dow1,dow2,dow3,dow4...dow6
    1       0    0    1    2    2      0
  select 
  user_id,
  sum(case order_dow when '0' then 1 else 0 end) as dow0,
  sum(case order_dow when '1' then 1 else 0 end) as dow1,
  sum(case order_dow when '2' then 1 else 0 end) as dow2,
  sum(case order_dow when '3' then 1 else 0 end) as dow3,
  sum(case order_dow when '4' then 1 else 0 end) as dow4,
  sum(case order_dow when '5' then 1 else 0 end) as dow5,
  sum(case order_dow when '6' then 1 else 0 end) as dow6
  from orders
  group by user_id
  limit 10;

业务4：

每个用户平均每个购买天中，购买的商品数量是多少？

比如有一个user ，有两天是购买商品的 ：
day 1 ：两个订单，1订单：10个product，2订单：15个product ， 共25个
day 2：一个订单， 1订单：12个product 共12个
day 3： 没有购买 

这个user 平均每个购买天的商品数量：
(10+15+12)/2 ,这里是2天有购买行为，不是三天，也不是3个订单。

select ord.user_id,sum(pri.products_cnt)/ count(distinct ord.days_since_prior_order) as avg_prod  
--date: 2离第一个商品10天,3离第二个10天 
from 
(select order_id,
user_id, 
if(days_since_prior_order='','-1',days_since_prior_order) as days_since_prior_order 
from orders where eval_set='prior')ord 
join 
(select order_id,count(1) as products_cnt from priors group by order_id)pri 
on ord.order_id=pri.order_id
group by ord.user_id
limit 10; 

业务5：每个用户最喜爱购买的三个product是什么，最终表结构可以是三个列，也可以是一个字符串/数组

select user_id, 
collect_list(concat_ws('_',product_id,cast(row_num as string))) as pro_top3,
size(collect_list(product_id)) as top_size
from
(
select user_id,
product_id, 
row_number() over(partition by user_id order by top_cnt desc) as row_num
from
(
-- user_id，product_id,count(1)这个用户其中一个商品购买了几次
select 
ord.user_id as user_id,
pri.product_id as product_id,
count(1) over(partition by user_id,product_id) as top_cnt
from 
(select * from orders where eval_set='prior')ord
join 
(select * from priors limit 10000)pri 
on ord.order_id=pri.order_id
-- group by ord.user_id,pri.product_id
)t 
)t1  where row_num <=3 group by user_id limit 10;