Bootstrap

Oracle之开窗函数使用

开窗函数是数据开发面试的必知必会,必须认真对待,上难度:
开窗函数语法格式如下,一共五部分:

①函数(a)over(<partition by b><order by c><windowing_clause> ⑤开窗范围)

1、函数:聚合函数,sum(a)、avg(a)、lag(a,n,null)、lead(a,n,null)、max(a)等
2、分组:partition by b
3、排序:order by c
4、开窗子句:range、rows、Specifying
range:排序字段有相同值时,开窗结果相同
rows:排序字段有相同值时,开窗结果不同
使用开窗子句一定要有排序!!
5、开窗范围:between…and…

  between m preceding and n following  --当前行的前m行到当前行的后n行
  between unbounded preceding and current row --第一行到当前行

–unbounded:不受控制的,无限的
–preceding:在…之前
–following:在…之后
–current row:当前行

以上为理论,以下为实际案例
————————————————-————

1、简单求和

sum(a)

2、累加和

sum(a)over(order by b)

3、分组求和

sum(a)over(partition by c order by b)

4、第一行到当前行分组求和(排序数据重复时,分析函数数据重复)

sum(a)over(partition by c order by b range between unbounded 
preceding and current row)

5、第一行到当前行分组求和(排序数据重复时,分析函数数据不重复)

sum(a)over (partition by c order by b rows between unbounded preceding and current row)

6、第一行到当前行前一行求和

sum(a)over(partition by c order by b range between unbounded
preceding and 1 preceding)

7、第一行到当前行后一行求和

sum(a)over(partition by c order by b range between unbounded preceding and 1 following)

8、当前行前80行到当前行求和

sum(a)over(partition by c order by b range between 79 preceding and current row)
;