wm_concat在行转列的时候非常有用，但在行转列的过程中的排序问题常常难以控制。
可见下面例子：
准备测试表：
drop table t;
create table t (n number,m number);
insert into t values(1,1);
insert into t values(5,3);
insert into t values(3,3);
insert into t values(6,5);
insert into t values(7,2);
insert into t values(2,2);
insert into t values(0,1);
insert into t values(11,1);
insert into t values(15,3);
insert into t values(13,3);
insert into t values(16,5);
insert into t values(17,2);
insert into t values(12,2);
insert into t values(10,1);
commit;
sql> select * from t order by 2,1;
N M
———- ———-
0 1
1 1
10 1
11 1
2 2
7 2
12 2
17 2
3 3
5 3
13 3
15 3
6 5
16 5
--按照m分组字段，n排序字段取出第一行
select * from ( select m,row_number() over (partition by m order by n) rn from t ) where rn<=1
