如何获得每天最多参与的ID |甲骨文

Question

我有点困惑如何为某些场景编写查询

表：

employee

create table employee
(
   name varchar2(10),
   sdate date,
   subid number
);

insert into dummy values ('Arun',to_date('2016-03-01','YYYY-MM-DD'),123);
insert into dummy values ('Arun',to_date('2016-03-01','YYYY-MM-DD'),453);
insert into dummy values ('Raj',to_date('2016-03-01','YYYY-MM-DD'),12);
insert into dummy values ('Raj',to_date('2016-03-01','YYYY-MM-DD'),45);
insert into dummy values ('Raj',to_date('2016-03-01','YYYY-MM-DD'),16);
insert into dummy values ('Raj',to_date('2016-03-01','YYYY-MM-DD'),18);
insert into dummy values ('Darshan',to_date('2016-03-01','YYYY-MM-DD'),1600);
insert into dummy values ('Darshan',to_date('2016-03-01','YYYY-MM-DD'),1820);

下面是我表格中的数据

name,date,subid
Arun,2016-03-01,123
Arun,2016-03-01,453
Raj,2016-03-01,12
Raj,2016-03-01,45
Raj,2016-03-01,16
Raj,2016-03-01,18
Darshan,2016-03-01,1600
Darshan,2016-03-01,1820

现在我需要以这样的方式进行过滤，让

maximum time

subid

参与

name

each day

Arun    -> count -> 2  -> 2016-03-01
Raj     -> count -> 4  -> 2016-03-01
Darshan -> count -> 2  -> 2016-03-01

输出：

name , subid , date
Raj  , 12    , 2016-03-01   

假设如果出现类似

Raj

和

Darshan

作为

same count

那么

order name wise

和

pick lesser subid

的名称

name,date,subid
Arun,2016-03-01,123
Arun,2016-03-01,453
Raj,2016-03-01,12
Raj,2016-03-01,45
Raj,2016-03-01,16
Raj,2016-03-01,18
Darshan,2016-03-01,1600
Darshan,2016-03-01,1820
Darshan,2016-03-01,160
Darshan,2016-03-01,18

Arun    -> count -> 2  -> 2016-03-01
Raj     -> count -> 4  -> 2016-03-01
Darshan -> count -> 4  -> 2016-03-01

字母顺序

Arun    -> count -> 2  -> 2016-03-01
Darshan -> count -> 4  -> 2016-03-01
Raj     -> count -> 4  -> 2016-03-01

选择较小的补贴

Darshan

输出：

name      , subid , date
Darshan   , 18    , 2016-03-01

我无法确定如何实现它。

我想到做

partition

by

，在什么基础上需要

partition

？

Answer 1

您似乎想要

GROUP BY

name

和日

TRUNC(dt)

，然后找到每个组的

MIN

imum

subid

，然后，对于计数最多的组（如果存在平局，则名字按字母顺序排列），

FETCH

FIRST ROW

：

SELECT name,
       TRUNC(dt) AS dt,
       MIN(subid) AS subid
FROM   table_name
GROUP BY name, TRUNC(dt)
ORDER BY COUNT(*) DESC, name
FETCH FIRST ROW ONLY