sql常见开窗函数示例

sql常见开窗函数示例 ​

1、开窗函数基本语法：

　　函数名()over(partition by 分区字段 order by 排序字段)

2.1、示例数据：

2.2、示例sql：

create table kchs(
ID int,
product varchar(50),
amount decimal(18,2)
)

insert into kchs values (1,'苹果',100);
insert into kchs values (2,'苹果',200);
insert into kchs values (3,'苹果',300);
insert into kchs values (4,'香蕉',450);
insert into kchs values (5,'香蕉',550);
insert into kchs values (6,'香蕉',650);
insert into kchs values (7,'西瓜',750);
insert into kchs values (8,'西瓜',850);
insert into kchs values (9,'西瓜',950);
insert into kchs values (10,'西瓜',950);
insert into kchs values (11,'西瓜',1050);

-- sum开窗，用于累计求和
select *,sum(amount)over(partition by product order by id) as sum_over from kchs order by id;

-- count开窗，用于统计当前分区当前行及以前行的纪录数
select *,count(*)over(partition by product order by id) as count_over from kchs order by id;

-- max开窗，用于统计当前分区当前行及以前行的最大值
select *,max(amount)over(partition by product order by id) as max_over from kchs order by id;

-- min开窗，用于统计当前分区当前行及以前行的最小值
select *,min(amount)over(partition by product order by id) as min_over from kchs order by id;

-- avg开窗，用于统计当前分区当前行及以前行的平均数
select *,avg(amount)over(partition by product order by id) as avg_over from kchs order by id;

-- lag开窗，用于获取往前偏移N行的数据
select *,lag(amount,1,null)over(partition by product order by id) as lag_over from kchs order by id;

-- lead开窗，用于获取往后偏移N行的数据
select *,lead(amount,1,null)over(partition by product order by id) as lead_over from kchs order by id;

-- rank开窗，用于计算当前分区按排序规则的排名，会并列排名，排名不连续，比如并列第三名有两个，那么就不会有第四名
select *,rank()over(partition by product order by amount) as rk from kchs order by id;

-- dense_rank开窗，用于计算当前分区按排序规则的排名，会并列排名，排名连续，比如并列第三名有两个，那么仍然会有第四名
select *,dense_rank()over(partition by product order by amount) as drk from kchs order by id;

-- row_number开窗，用于计算当前分区按排序规则的排序，不会并列排序
select *,row_number()over(partition by product order by amount) as rn from kchs order by id;

View Code

3.1、sum开窗：

-- sum开窗，用于累计求和
select *,sum(amount)over(partition by product order by id) as sum_over from kchs order by id;

3.2、count开窗：

-- count开窗，用于统计当前分区当前行及以前行的纪录数
select *,count(*)over(partition by product order by id) as count_over from kchs order by id;

 3.3、max开窗：

-- max开窗，用于统计当前分区当前行及以前行的最大值
select *,max(amount)over(partition by product order by id) as max_over from kchs order by id;

 3.4、min开窗：

-- min开窗，用于统计当前分区当前行及以前行的最小值
select *,min(amount)over(partition by product order by id) as min_over from kchs order by id;

 3.5、avg开窗：

-- avg开窗，用于统计当前分区当前行及以前行的平均数
select *,avg(amount)over(partition by product order by id) as avg_over from kchs order by id;

 3.6、lag开窗：

-- lag开窗，用于获取往前偏移N行的数据
select *,lag(amount,1,null)over(partition by product order by id) as lag_over from kchs order by id;

 3.7、lead开窗：

-- lead开窗，用于获取往后偏移N行的数据
select *,lead(amount,1,null)over(partition by product order by id) as lead_over from kchs order by id;

3.8、rank开窗：

-- rank开窗，用于计算当前分区按排序规则的排名，会并列排名，排名不连续，比如并列第三名有两个，那么就不会有第四名
select *,rank()over(partition by product order by amount) as rk from kchs order by id;

 3.9、dense_rank开窗：

-- dense_rank开窗，用于计算当前分区按排序规则的排名，会并列排名，排名连续，比如并列第三名有两个，那么仍然会有第四名
select *,dense_rank()over(partition by product order by amount) as drk from kchs order by id;

3.10、row_number开窗：

-- row_number开窗，用于计算当前分区按排序规则的排序，不会并列排序
select *,row_number()over(partition by product order by amount) as rn from kchs order by id;