高级sql

窗口函数

CREATE TABLE students ( id INT PRIMARY KEY auto_increment, NAME VARCHAR ( 30 ), math INT, english INT, num INT );
INSERT INTO students ( NAME, math, english, num )
VALUES
	( "张三", 99, 88, 4 ),
	( "李四", 77, 88, 5 ),
	( "王五", 96, 74, 3 ),
	( "赵六", 76, 89, 1 ),
	( "天好", 85, 81, 2 );

SELECT
	id,
	math,
	lag( math ) over ( ORDER BY math ) AS lag1,
	lag( math, 1 ) over ( ORDER BY math ) AS lag2,
	lag( math, 1, NULL ) over ( ORDER BY math ) AS lag3,
	lead( math ) over ( ORDER BY math ) AS lead1,
	class,
	FIRST_VALUE( math ) over ( PARTITION BY class ORDER BY id ASC rows BETWEEN unbounded preceding AND unbounded following ) AS FIRST,
	LAST_VALUE( math ) over ( PARTITION BY class ORDER BY id ASC rows BETWEEN unbounded preceding AND unbounded following ) AS last,
	sum( math ) over ( ORDER BY math ) AS sum1,
	sum( math ) over ( ORDER BY math rows BETWEEN unbounded preceding AND current ROW ) AS sum2,
	sum( math ) over ( PARTITION BY class ) AS sum3,
	sum( math ) over ( PARTITION BY class rows BETWEEN unbounded preceding AND unbounded following ) AS sum4,
	sum( math ) over ( PARTITION BY class ORDER BY math ) AS sum5,
	sum( math ) over ( PARTITION BY class ORDER BY math rows BETWEEN unbounded preceding AND current ROW ) AS sum6,
	sum( math ) over ( ORDER BY math RANGE BETWEEN 10 preceding AND 10 following ) AS sum7,
	rank() over ( PARTITION BY class ORDER BY math ) AS rank1,
	dense_rank() over ( PARTITION BY class ORDER BY math ) AS rank2,
	row_number() over ( PARTITION BY class ) AS row_num 
FROM
	students;

• lag(math): 窗口函数，需要搭配 over 使用，有三个参数，后两个参数可以省略。
• lag(math) over (ORDER BY math) : 按数学成绩升序，得到之前一行的同学的数学成绩
• lag(math, 1) over (ORDER BY math): 同上，参数 1 表示错位一行，即上一行，默认为1，可以省略，
• lag(math, 1, NULL) over (ORDER BY math): 同上，参数null表示若不存在前一行，该值为null，默认为null，可以省略
• lead(math) over (ORDER BY math)：按数学成绩升序排序，得到后一行的同学的数学成绩。
• FIRST_VALUE(math): 获取指定列math在窗口范围内的第一行的值
• over ( )：指定窗口函数的窗口范围。
  • partition by class: 按班级分组
  • order by math: 窗口内行记录的排序规则，在窗口内按数学成绩排序
  • rows BETWEEN unbounded preceding AND unbounded following : 限制窗口的边界，这里表示窗口范围从当前行之前的第一行（无限制）到当前行之后的最后一行（无限制）
• FIRST_VALUE( math ) over ( PARTITION BY class ORDER BY id ASC rows BETWEEN unbounded preceding AND unbounded following ) AS FIRST : 在每个班级内，学生按照id排序，获取排序第一的学生的数学成绩。这样，每一行都会有一个额外的列"FIRST"，其中包含了每个班级内第一个学生的数学成绩。
• LAST_VALUE( math ): 用于获取指定列math在窗口范围内的最后一个值。
• sum( math ) over ( ORDER BY math ): 窗口按数学成绩排序，获取排序后当前行及之前所有行的数学成绩总和。
• sum( math ) over ( ORDER BY math rows BETWEEN unbounded preceding AND current ROW ) : 同上，显示指定了滑动窗口范围是 排序后的当前行及之前所有行
• sum( math ) over ( PARTITION BY class ): 获取每个班的所有学生数学成绩的总和
• rank() over ( PARTITION BY class ORDER BY math ) : 获取每个班内的数学成绩排名
  • 排名相同的行会被分配相同的排名值，而下一行就会跳过这些排名值
  • 如 1、1、3、4、5、5、5、8
• dense_rank():也是窗口函数，需要搭配over使用
  • 允许有相同的排名
  • 排名值整体连续
  • 如1、2、2、3、4、5、6、7、7、7、8
• row_number: 也是窗口函数，需要搭配over使用，给窗口内的数据分配行号
  • 行号依次递增
  • 如果两个行排序字段相同，行号也是依次递增
  • 如：1、2、3、5、6、7、8、9
• sum(math) over (ORDER BY math RANGE BETWEEN 10 preceding AND 10 following):
  • 按数学成绩排序
  • RANGE BETWEEN 10 preceding AND 10 following: 当前行的窗口范围是 数学成绩不低于当前行十分的且不高于当前行十分的所有学生。
  • range 表示按照数值选取范围
  • rows:表示按照行来选取范围

查询结果排序

select * from students order by math;

• 对结果集进行排序

select * from students distribute by class sort by math;

• distribute by： 指定class作为分发键，保证了相同的class一定会分发给相同的producer处理
• sort by：在单个producer内进行排序的规则，能保证局部有序，不保证全局有序。

select * from students cluster by math;

• cluster by math 等价于 distribute by math sort by math