MySQL - 全表分组后，获取组内排序首条数据信息

news2024/10/7 12:19:48

性能
- 不详!!! 不详!!! 不详!!! 请谨慎使用!!!
环境
- MySQL服务: 8.0+版本;
思路
- 使用8.0+版本的新函数特性:
  - row_number(): 序号函数; 顾名思义, 就是给每组中的元素从1开始按顺序加上序号;
  - over(): 其中两个语法如下
    - partition: 按某字段分组;
    - order by: 按某字段排序;
  - 注意: 两函数详细使用方法可自行查询;
第一步建表:

CREATE TABLE `t_score`  (
  `id` int NOT NULL AUTO_INCREMENT,
  `name` varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NULL DEFAULT NULL COMMENT '姓名',
  `sex` varchar(10) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NULL DEFAULT NULL COMMENT '性别',
  `score` double(20, 2) NULL DEFAULT NULL COMMENT '成绩',
  `subject` varchar(255) CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci NULL DEFAULT NULL COMMENT '科目',
  `create_time` datetime NULL DEFAULT NULL,
  `update_time` datetime NULL DEFAULT NULL,
  PRIMARY KEY (`id`) USING BTREE
) ENGINE = InnoDB CHARACTER SET = utf8mb4 COLLATE = utf8mb4_0900_ai_ci ROW_FORMAT = Dynamic;

第二步加数据:

INSERT INTO `t_score` VALUES (1, '张三', '男', 63.00, '语文', '2023-09-06 10:06:22', '2023-09-26 10:06:25');
INSERT INTO `t_score` VALUES (2, '李四', '男', 75.00, '语文', '2023-09-26 10:06:32', '2023-09-26 10:06:36');
INSERT INTO `t_score` VALUES (3, '小美', '女', 89.00, '语文', '2023-09-26 10:06:46', '2023-09-26 10:06:48');
INSERT INTO `t_score` VALUES (4, '张三', '男', 78.00, '数学', '2023-09-06 10:06:22', '2023-09-26 10:06:25');
INSERT INTO `t_score` VALUES (5, '李四', '男', 79.00, '数学', '2023-09-26 10:06:32', '2023-09-26 10:06:36');
INSERT INTO `t_score` VALUES (6, '小美', '女', 94.00, '数学', '2023-09-26 10:06:46', '2023-09-26 10:06:48');
INSERT INTO `t_score` VALUES (7, '张三', '男', 45.00, '英语', '2023-09-06 10:06:22', '2023-09-26 10:06:25');
INSERT INTO `t_score` VALUES (8, '李四', '男', 34.00, '英语', '2023-09-26 10:06:32', '2023-09-26 10:06:36');
INSERT INTO `t_score` VALUES (9, '小美', '女', 99.00, '英语', '2023-09-26 10:06:46', '2023-09-26 10:06:48');

第三步看两函数效果:
- 需求: 把每个人的成绩, 按照从高到低的顺序加上序号;
- 思路: 按名称分组, 成绩排序, 然后给每组各行数据加上序号;
- ```
SELECT 
	id, 
	`name`, 
	`score`,
	`subject`, 
	ROW_NUMBER()OVER(PARTITION BY `name` ORDER BY `SCORE` DESC) AS row_index 
FROM t_score;
```
  效果:

第四部实现最终需求:

思路: 按名称分组,成绩排序,然后获取序号为1的数据;

脚本:

SELECT * 
FROM(
	SELECT 
		id, 
		`name`, 
		`score`,
		`subject`, 
		ROW_NUMBER()OVER(PARTITION BY `name` ORDER BY `SCORE` DESC) AS row_index 
	FROM t_score) AS ib
WHERE ib.row_index = 1;