mfy学习笔记

创建表并导入数据

CREATE TABLE sales (
    id INT,
    salesperson STRING,
    region STRING,
    sales_amount INT,
    sale_date DATE
);

INSERT INTO sales (id, salesperson, region, sales_amount, sale_date)
VALUES
(1, 'Alice', 'North', 1000, '2023-01-01'),
(2, 'Bob', 'South', 1500, '2023-01-02'),
(3, 'Alice', 'North', 2000, '2023-01-03'),
(4, 'Charlie', 'East', 1200, '2023-01-04'),
(5, 'Bob', 'South', 1800, '2023-01-05'),
(6, 'Alice', 'North', 2500, '2023-01-06'),
(7, 'Charlie', 'East', 1300, '2023-01-07'),
(8, 'Bob', 'South', 2200, '2023-01-08'),
(9, 'Alice', 'North', 3000, '2023-01-09'),
(10, 'Charlie', 'East', 1400, '2023-01-10');

示例数据表：sales

id	salesperson	region	sales_amount	sale_date
1	Alice	North	1000	2023-01-01
2	Bob	South	1500	2023-01-02
3	Alice	North	2000	2023-01-03
4	Charlie	East	1200	2023-01-04
5	Bob	South	1800	2023-01-05
6	Alice	North	2500	2023-01-06
7	Charlie	East	1300	2023-01-07
8	Bob	South	2200	2023-01-08
9	Alice	North	3000	2023-01-09
10	Charlie	East	1400	2023-01-10

---

### 1. `CASE WHEN` 示例
根据销售额给销售人员分类：

```sql
SELECT 
    salesperson,
    sales_amount,
    CASE 
        WHEN sales_amount < 1500 THEN 'Low'
        WHEN sales_amount BETWEEN 1500 AND 2500 THEN 'Medium'
        ELSE 'High'
    END AS sales_category
FROM 
    sales;

运行结果：

salesperson	sales_amount	sales_category
Alice	1000	Low
Bob	1500	Medium
Alice	2000	Medium
Charlie	1200	Low
Bob	1800	Medium
Alice	2500	Medium
Charlie	1300	Low
Bob	2200	Medium
Alice	3000	High
Charlie	1400	Low

---

2. SUM(CASE WHEN) 示例

计算每个区域的销售总额：

SELECT 
    region,
    SUM(CASE WHEN salesperson = 'Alice' THEN sales_amount ELSE 0 END) AS alice_sales,
    SUM(CASE WHEN salesperson = 'Bob' THEN sales_amount ELSE 0 END) AS bob_sales,
    SUM(CASE WHEN salesperson = 'Charlie' THEN sales_amount ELSE 0 END) AS charlie_sales
FROM 
    sales
GROUP BY 
    region;

运行结果：

region	alice_sales	bob_sales	charlie_sales
North	8500	0	0
South	0	5500	0
East	0	0	3900

---

3. RANK() 示例

根据销售额对销售人员进行排名：

SELECT 
    salesperson,
    sales_amount,
    RANK() OVER (ORDER BY sales_amount DESC) AS sales_rank
FROM 
    sales;

运行结果：

salesperson	sales_amount	sales_rank
Alice	3000	1
Bob	2200	2
Alice	2500	3
Bob	1800	4
Alice	2000	5
Bob	1500	6
Charlie	1400	7
Charlie	1300	8
Alice	1000	9
Charlie	1200	10

---

4. ROW_NUMBER() 示例

为每个销售人员的销售额分配一个唯一的行号：

SELECT 
    salesperson,
    sales_amount,
    ROW_NUMBER() OVER (PARTITION BY salesperson ORDER BY sales_amount DESC) AS row_num
FROM 
    sales;

运行结果：

salesperson	sales_amount	row_num
Alice	3000	1
Alice	2500	2
Alice	2000	3
Alice	1000	4
Bob	2200	1
Bob	1800	2
Bob	1500	3
Charlie	1400	1
Charlie	1300	2
Charlie	1200	3

---

5. DENSE_RANK() 示例

根据销售额对销售人员进行密集排名（不会跳过排名）：

SELECT 
    salesperson,
    sales_amount,
    DENSE_RANK() OVER (ORDER BY sales_amount DESC) AS dense_rank
FROM 
    sales;

运行结果：

salesperson	sales_amount	dense_rank
Alice	3000	1
Bob	2200	2
Alice	2500	3
Bob	1800	4
Alice	2000	5
Bob	1500	6
Charlie	1400	7
Charlie	1300	8
Alice	1000	9
Charlie	1200	10

---

总结

• CASE WHEN：用于条件判断，生成新的列。
• SUM(CASE WHEN)：用于按条件汇总数据。
• RANK()：用于排名，允许并列排名并跳过后续名次。
• ROW_NUMBER()：用于生成唯一的行号，即使数据相同也会分配不同行号。
• DENSE_RANK()：用于密集排名，允许并列排名但不跳过后续名次。

通过这些示例和运行结果，可以清晰地展示每个函数的作用和用法！