我有一张表格,其中记录了不定期的产品价格:
prod_id | price | timestamp
p1 | 10 | 01-01-2018
p1 | 25 | 03-01-2018
p1 | 15 | 05-01-2018
p2 | 40 | 03-01-2018
p2 | 80 | 04-01-2018
p2 | 90 | 08-01-2018
p3 | 150 | 03-01-2018
p3 | 200 | 05-01-2018
给定一个时间范围和一个阈值,我想查询该表并获取其时间周期内的百分比增长大于输入阈值的产品。
例如,如果时间范围是从01-01-2018到06-01-2018,则p1,p2和p3的增加百分比为:
p1 -> 50%
p2 -> 100%
p3 -> 33.33%
因此,如果阈值是40%,则响应应该是
prod_id |
p1
p2
是否可以使用SQL做到这一点?
答案 0 :(得分:1)
尝试一下:
样本数据:
create table tbl (prod_id char(2), price int, `timestamp` date);
insert into tbl values
('p1', 10, '2018-01-01'),
('p1', 25, '2018-01-03'),
('p1', 15, '2018-01-05'),
('p2', 40, '2018-01-03'),
('p2', 80, '2018-01-04'),
('p2', 90, '2018-01-08'),
('p3', 150, '2018-01-03'),
('p3', 200, '2018-01-05');
T-SQL:
set @start = '2018-01-01', @end = '2018-01-06', @threshold = 0.4;
select a.prod_id, (c.price - b.price)/b.price from (
select prod_id, min(`timestamp`) `min`, max(`timestamp`) `max` from tbl
where `timestamp` between @start and @end
group by prod_id
) a join tbl b on a.prod_id = b.prod_id and a.`min` = b.`timestamp`
join tbl c on a.prod_id = c.prod_id and a.`max` = c.`timestamp`
where (c.price - b.price)/b.price > @threshold;