文章目录
- 测试数据
- 需求说明
- 需求实现
- 分步解析
测试数据
-- 创建商品表
DROP TABLE IF EXISTS products;
CREATE TABLE products (product_id INT,product_name STRING
);INSERT INTO products VALUES
(1, 'Product A'),
(2, 'Product B'),
(3, 'Product C'),
(4, 'Product D'),
(5, 'Product E'),
(6, 'Product F'),
(7, 'Product G'),
(8, 'Product H'),
(9, 'Product I'),
(10, 'Product J'),
(11, 'Product K');-- 创建销售表
DROP TABLE IF EXISTS sales;
CREATE TABLE sales (sale_id INT,product_id INT,sale_date STRING,amount DOUBLE
);INSERT INTO sales VALUES
(101, 1, '2023-01-01', 100.0),
(102, 1, '2023-02-01', 150.0),
(103, 2, '2023-03-01', 200.0),
(104, 3, '2023-04-01', 50.0),
(105, 4, '2023-05-01', 300.0),
(106, 5, '2023-06-01', 250.0),
(107, 1, '2024-01-01', 120.0),
(108, 1, '2024-02-01', 180.0),
(109, 2, '2024-03-01', 220.0),
(110, 3, '2024-04-01', 70.0),
(111, 4, '2024-05-01', 330.0),
(112, 5, '2024-06-01', 270.0),
(113, 2, '2023-07-01', 180.0),
(114, 3, '2023-08-01', 60.0),
(115, 4, '2023-09-01', 310.0),
(116, 5, '2023-10-01', 260.0),
(117, 1, '2023-11-01', 190.0),
(118, 2, '2023-12-01', 210.0),
(119, 3, '2024-01-01', 75.0),
(120, 4, '2024-02-01', 340.0),
(121, 5, '2024-03-01', 280.0),
(122, 6, '2023-01-01', 130.0),
(123, 6, '2023-02-01', 160.0),
(124, 7, '2023-03-01', 190.0),
(125, 8, '2023-04-01', 220.0),
(126, 9, '2023-05-01', 250.0),
(127, 10, '2023-06-01', 280.0),
(128, 6, '2024-01-01', 140.0),
(129, 6, '2024-02-01', 170.0),
(130, 7, '2024-03-01', 200.0),
(131, 8, '2024-04-01', 230.0),
(132, 9, '2024-05-01', 260.0),
(133, 10, '2024-06-01', 290.0),
(134, 7, '2023-07-01', 175.0),
(135, 8, '2023-08-01', 205.0),
(136, 9, '2023-09-01', 235.0),
(137, 10, '2023-10-01', 265.0),
(138, 6, '2023-11-01', 145.0),
(139, 7, '2023-12-01', 175.0),
(140, 8, '2024-01-01', 215.0),
(141, 9, '2024-02-01', 245.0),
(142, 10, '2024-03-01', 275.0),
(143, 6, '2024-04-01', 155.0),
(144, 7, '2024-05-01', 185.0),
(145, 8, '2024-06-01', 225.0),
(147, 11, '2023-06-09', 0.0),
(146, 11, '2024-06-01', 233.0);
需求说明
统计各个商品今年销售额与去年销售额的增长率及销售额的排名变化。
增长率计算公式:(当期份额-上期份额)/ 上期份额 * 100%
结果示例:
product_name | total_amount_2023 | total_amount_2024 | growth_rate | rk_2023 | rk_2024 | rk_diff |
---|---|---|---|---|---|---|
Product D | 610.0 | 670.0 | 9.8% | 1 | 1 | 0 |
Product H | 425.0 | 670.0 | 57.6% | 9 | 1 | 8 |
Product J | 545.0 | 565.0 | 3.7% | 3 | 3 | 0 |
Product E | 510.0 | 550.0 | 7.8% | 5 | 4 | 1 |
Product I | 485.0 | 505.0 | 4.1% | 6 | 5 | 1 |
… | … | … | … | … | … | … |
其中:
product_name
表示商品名称;total_amount_2023
表示商品在2023
年度的销售额;total_amount_2024
表示商品在2024
年度的销售额;growth_rate
表示商品的增长率;rk_2023
表示商品在2023
年度中的销售额排名;rk_2024
表示商品在2024
年度中的销售额排名;rk_diff
表示该商品年度销售额排名的变化。
注意,在这里商品销售额可能存在两种情况:
- 假设某商品 2023 年销售
0.0
,而在2024
年销售50
,那么这种情况下,销售额增长率统一设置为100.0%
;- 如果在两个年度销售均为
0.0
,那么销售额增长率设置为0.0%
。
需求实现
SELECTp.product_name,total_amount_2023,total_amount_2024,CASE WHEN total_amount_2024=0 AND total_amount_2023=0THEN "0.0%"WHEN total_amount_2023=0THEN "100.0%"ELSECONCAT(CAST((total_amount_2024 - total_amount_2023) / total_amount_2023 as DECIMAL(5,3)) * 100,"%")END growth_rate,rk_2023,rk_2024,rk_2024 - rk_2023 rk_diff
FROM(SELECTproduct_id,total_amount_2023,total_amount_2024,RANK() OVER(ORDER BY total_amount_2023 DESC) rk_2023,RANK() OVER(ORDER BY total_amount_2024 DESC) rk_2024FROM(SELECTproduct_id,SUM(IF(year(sale_date)="2023",amount,0)) total_amount_2023,SUM(IF(year(sale_date)="2024",amount,0)) total_amount_2024FROMsalesWHEREyear(sale_date) IN ("2023","2024")GROUP BYproduct_id)t1 )t2
JOINproducts p
ONt2.product_id = p.product_id;
输出结果如下:
分步解析
(1)获取去年与今年两个年度的数据,并进行聚合统计。
SELECTproduct_id,SUM(IF(year(sale_date)="2023",amount,0)) total_amount_2023,SUM(IF(year(sale_date)="2024",amount,0)) total_amount_2024
FROMsales
WHEREyear(sale_date) IN ("2023","2024")
GROUP BYproduct_id;
(2)根据(1)中的结果,通过窗口函数排序,获取分别获取两个年度的销售额排名。
SELECTproduct_id,total_amount_2023,total_amount_2024,RANK() OVER(ORDER BY total_amount_2023 DESC) rk_2023,RANK() OVER(ORDER BY total_amount_2024 DESC) rk_2024
FROM(SELECTproduct_id,SUM(IF(year(sale_date)="2023",amount,0)) total_amount_2023,SUM(IF(year(sale_date)="2024",amount,0)) total_amount_2024FROMsalesWHEREyear(sale_date) IN ("2023","2024")GROUP BYproduct_id)t1;
(3)根据(2)中的结果,判断并计算两个年度的增长率以及排名变化,最终通过 join
连接商品表,获取商品名称。
SELECTp.product_name,total_amount_2023,total_amount_2024,CASE WHEN total_amount_2024=0 AND total_amount_2023=0THEN "0.0%"WHEN total_amount_2023=0THEN "100.0%"ELSECONCAT(CAST((total_amount_2024 - total_amount_2023) / total_amount_2023 as DECIMAL(5,3)) * 100,"%")END growth_rate,rk_2023,rk_2024,rk_2023 - rk_2024 rk_diff
FROM(SELECTproduct_id,total_amount_2023,total_amount_2024,RANK() OVER(ORDER BY total_amount_2023 DESC) rk_2023,RANK() OVER(ORDER BY total_amount_2024 DESC) rk_2024FROM(SELECTproduct_id,SUM(IF(year(sale_date)="2023",amount,0)) total_amount_2023,SUM(IF(year(sale_date)="2024",amount,0)) total_amount_2024FROMsalesWHEREyear(sale_date) IN ("2023","2024")GROUP BYproduct_id)t1 )t2
JOINproducts p
ONt2.product_id = p.product_id;
可能对于排名那里存在疑惑,为什么是 rk_2023 - rk_2024
,不是 rk_2024 - rk_2023
呢?
惯性思维导致,在排序中,并不是排名越高值越大,相反,因为我们的排名越靠前(越高),其排名值越小,想到这里,就应该明白了。