db2中关于递归(with)的使用--合并行列

db2中关于递归(with)的使用

因为有人问及，这边简单的再探讨一下
-------------------
1.创建测试表
create table zxt_test
( id varchar(10),
  ivalue varchar(20),
  iname varchar(20)
)
commit;
-----------
2.插入测试语句
insert into zxt_test values('1','aa','x'),('2','bb','x'),('3','bb','x'),('1','bb','y'),('2','bb','y'),('3','bb','y');
commit;
---------------
3.查看数据
select * from zxt_test;
1    'aa'    'x'
2    'bb'    'x'
3    'bb'    'x'
1    'bb'    'y'
2    'bb'    'y'
3    'bb'    'y'
----------------
4.with 的写法！
with
s as (
select row_number()over(partition by iname order by id) id1,
       row_number()over(partition by iname order by id) id2, ivalue,iname from zxt_test
)
,
t(iname,id1,id2,ivalue) as
(
select iname,id1,id2,cast(ivalue as varchar(100)) from  s where id1 =1 and id2=1
union all
select t.iname,t.id1+1,t.id2,cast(s.ivalue||','||t.ivalue as varchar(100))
from  s, t
where   s.id2=t.id1+1 and t.iname = s.iname
)
select iname,ivalue from t where t.id1= (select max(id1) from s where s.iname = t.iname);
5.结果如下：
'x'    'bb,bb,aa'
'y'    'bb,bb,bb'