cume_dist()：
功能描述：計算一行在組中的相對位置，返回大於0 小於等於1的數，例如在乙個4行組中，分布是1/4 ,2/4 ,3/4 ，4/4
如果有重複的就乘以重複的個數，如第一行和第二行重複 分布為 (1/4)*2,(1/4)*2, 3/4 ,4/4
下面是例子：
表 all_sales
select
* from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
****************************************=
1 2003 1 1 21 10034.84
2 2003 1 2 21 1034.84
3 2003 1 3 21 1034.84
4 2003 1 4 21 3034.84
執行sql:
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount asc) as cume_dist
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
******************************==
1 2 1034.84 0.5
2 3 1034.84 0.5
3 4 3034.84 0.75
4 1 10034.84 1
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount asc) as cume_dist
,percent_rank() over(order by t.amount desc )as percent_rank
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
******************************===
1 1 10034.84 1 0
2 4 3034.84 0.75 0.333333333333333
3 2 1034.84 0.5 0.666666666666667
4 3 1034.84 0.5 0.666666666666667
ntilte()函式記錄n分片的值。
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount asc) as cume_dist
,percent_rank() over(order by t.amount asc )as percent_rank
,ntile(5) over(order by t.amount desc )as ntile
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
****************************************=
1 1 10034.84 1 1 1
2 4 3034.84 0.75 0.666666666666667 2
3 2 1034.84 0.5 0 3
4 3 1034.84 0.5 0 4
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount asc) as cume_dist
,percent_rank() over(order by t.amount asc )as percent_rank
,ntile(3) over(order by t.amount desc )as ntile
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
******************************
1 1 10034.84 1 1 1
2 4 3034.84 0.75 0.666667 1
3 2 1034.84 0.5 0 2
4 3 1034.84 0.5 0 3
row_number():
從1開始為每個分組返回乙個數字
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount asc) as cume_dist
,percent_rank() over(order by t.amount asc )as percent_rank
,ntile(2) over(order by t.amount desc )as ntile
,row_number() over(order by t.amount desc) as row_number
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
*************************====
1 1 10034.84 1 1 1 1
2 4 3034.84 0.75 0.666667 1 2
3 2 1034.84 0.5 0 2 3
4 3 1034.84 0.5 0 2 4
和percentile_cont(x)
的作用與cume_dist()
和percent_rank()
相反:
select
percentile_disc(0.25) within group(order by t.amount desc)
as percentile_disc
,percentile_cont(0.666666666666667) within group(order by t.amount desc)
as percentile_cont
from all_sales t
where year =2003
and t.month =1
and t.amount is not null
and t.emp_id=21
******************************====
1 10034.84 1034.84
select
t.prd_type_id
,t.amount
,cume_dist() over (order by t.amount desc) as cume_dist
,percent_rank() over(order by t.amount desc )as percent_rank
from all_sales t
where t.year=2003
and t.month =1
and t.amount is not null
and t.emp_id=21
********************==
1 1 10034.84 0.25 0 ***x u
2 4 3034.84 0.5 0.333333333333333 ***x u
3 3 1034.84 1 0.666666666666667 ***x u
4 2 1034.84 1 0.666666666666667 ***x u
orace後台執行
7.19 資料庫後台執行命令和客戶端登陸 sqlplus username password host port sid 普通使用者 sqlplus as sysdba或者conn as sysdba sys使用者 pl sql普通使用者登陸 database 10.171.37.11 1526 t...
orace遞迴查詢
查詢結果自己所有的前代節點 包括自己 select from orgtable o where nvl o.canceled,0 1 start with o.id 5661 connect by to char o.id prior to char o.supsubcomid 查詢結果自己所有的後...
orace時間往前推
oracle中獲取當前系統時間往前推7天的日期 2014 01 07 11 03 19 select sysdate interval 7 day from dual select to date to char sysdate,yyyy mm dd hh mi ss yyyy mm dd hh m...