删除重复项并模拟 group_concat 对结果进行排序

Question

我已经找到了这个有趣的线程

Simizing group_concat MySQL function在 Microsoft SQL Server 2005 中？

use test
go
create table methods (
id int identity,
id_exam int,
method int
)
go
insert into methods (id_exam,method) values (1,5)
insert into methods (id_exam,method) values (1,2)
insert into methods (id_exam,method) values (1,5)
insert into methods (id_exam,method) values (2,1)
insert into methods (id_exam,method) values (3,5)
insert into methods (id_exam,method) values (3,2)
insert into methods (id_exam,method) values (3,2)
insert into methods (id_exam,method) values (4,5)
insert into methods (id_exam,method) values (4,3)

select 
id_exam, 
method = replace ((select method AS [data()]
                   from methods
                   where id_exam = a.id_exam                      
                   order by id_exam for xml path('')), ' ', ',')
from  methods a
where id_exam is not null
group by id_exam

这给了我

1   5,2,5
2   1
3   5,2,2
4   5,3

但是我想从每次考试中删除重复项并对串联结果进行排序以获得

1  2,5
2  1
3  2,5
4  3,5

感谢。

Answer 1

尝试在内部查询中使用 DISTINCT 并按方法排序而不是 id_exam。

select 
id_exam, 
method = replace ((select distinct method AS [data()]
                   from methods
                   where id_exam = a.id_exam                      
                   order by method for xml path('')), ' ', ',')
from  methods a
where id_exam is not null
group by id_exam

Answer 2

将 group by method 添加到内部查询，并将 order by id_exam 更改为 order by method。

select 
id_exam, 
method = replace ((select method AS [data()]
                   from methods
                   where id_exam = a.id_exam
                   group by method                      
                   order by method for xml path('')), ' ', ',')
from  methods a
where id_exam is not null
group by id_exam

结果：

id_exam     method
----------- ---------
1           2,5
2           1
3           2,5
4           3,5