Oracel 行转列
需求：
课程表：tb_Course
| CID
|
CN(Course Name)
|
| 1
|
C1
|
| 2
|
C2
|
区域表：tb_Area
| AID
|
AN
|
| 1
|
A1
|
| 2
|
A2
|
| 3
|
A3
|
| 4
|
A4
|
| 5
|
A5
|
区域国家表: tb_Area_Country
| AID
|
AN(Area Name)
|
CoID (Country ID)
|
| 1
|
A1
|
1
|
| 1
|
A1
|
2
|
| 1
|
A1
|
3
|
| 2
|
A2
|
4
|
| 2
|
A2
|
5
|
| 2
|
A2
|
6
|
| 3
|
A3
|
7
|
| 4
|
A4
|
8
|
| 5
|
A5
|
9
|
课程到国家的同步状态 tb_record （没有记录的就是未同步）
| CID(Course ID)
|
CoID(Country ID)
|
Status
|
|
| 1
|
1
|
1
|
1:同步中,2:已同步
|
| 1
|
2
|
2
|
|
| 1
|
3
|
1
|
|
| 1
|
4
|
1
|
|
| 1
|
5
|
2
|
|
| 1
|
6
|
2
|
|
| 1
|
7
|
2
|
|
| 1
|
8
|
1
|
展示样例：
| CID
|
DistributeArea
(include tb_record.status is 2)
|
No_DistributeArea
(include tb_record.status is 0, 1)
|
| 1
|
A3(已同步)
|
A1(同步中)、A2(同步中)、A4(同步中)、A5(未同步)
|
| 2
|
A1(未同步)、A2(未同步)、A2(未同步)、A4(未同步)、A5(未同步)
|
实现：
第一步：先关联 tb_Area 和 tb_record 得到区域的同步状态AS以及区域分发状态AD：
Select tbR.CID, tbA.AN, decode(min(tbR.status),‘1’, ‘同步中’, ‘已同步’) AS, decode(min(tbR.status), ‘2’, ‘1’, ‘0’) AD
From tb_Area_Country tbA join tb_record tbR on tbA.CoID = tbR.CoID
Group by tbR.CID, tbA.AN 结果表：T1
| CID
|
AN
|
AS
|
AD
|
| 1
|
A1
|
同步中’
|
0
|
| 1
|
A2
|
同步中’
|
0
|
| 1
|
A3
|
已同步
|
1
|
| 1
|
A4
|
同步中’
|
0
|
第二步：用课程表和区域表做迪卡积，除去同步中及已同步的记录，再与同步中，已同步的记录做union，得到所有记录
Select c.cid, a.AN, '未同步' AS, '0' AD
from tb_Course c cross join tb_area a
where not exists(select 1 from
(Select tbR.CID, tbA.AN, decode(min(tbR.status),’1’,’同步中’,’已同步’) AS,
decode(min(tbR.status), ‘2’, ‘1’, ‘0’) AD
From tb_Area_Country tbA join tb_record tbR on tbA.CoID = tbR.CoID
Group by tbR.CID, tbA.AN
) t
Where t.cid = c.cid and t.AN = a.AN
)
union
Select tbR.CID, tbA.AN, decode(min(tbR.status),’1’,’同步中’,’已同步’) AS,
decode(min(tbR.status), ‘2’, ‘1’, ‘0’) AD
From tb_Area_Country tbA join tb_record tbR on tbA.CoID = tbR.CoID
Group by tbR.CID, tbA.AN;
| CID
|
AN
|
AS
|
AD
|
| 1
|
A1
|
同步中
|
0
|
| 1
|
A2
|
同步中
|
0
|
| 1
|
A3
|
已同步
|
1
|
| 1
|
A4
|
同步中
|
0
|
| 1
|
A5
|
未同步
|
0
|
| 2
|
A1
|
未同步
|
0
|
| 2
|
A2
|
未同步
|
0
|
| 2
|
A3
|
未同步
|
0
|
| 2
|
A4
|
未同步
|
0
|
| 2
|
A5
|
未同步
|
0
|
一二步可以合起来做，如果没有过滤条件可以用定义成视图，如果有查询条件就可以创建一个临时表，过滤后插入临时表。我们暂且定义为视图：T1.
第三步：思路是先构建一颗树，然后再将树转成列：
方法1：构建树，先设法构建父子关系表 T2
1、先给表排序获得行号作为父节点ID
Select T1.CID, T1.AN, T1.AS, T1.AD, row_number() over(order by T1.CID,T1.AD) rnFirst
From T1 得到T2
2、使用lead函数和分区特性得到子节点id
Select T2.CID, T2.AN, T2.AS, T2.AD, T2.rnFirst, lead(T2.rnFirst) over(partition by T2.CID,T2.AD order by T2.rnFirst) rnNext From T2 得到T3
3、构造树，利用sys_connect_by_path函数链接字符串
Select T3.CID, T3.AD, sys_connect_by_path(T3.AN ||’(‘ || T3.AS || ‘)’, ‘、’) Area
From T3
Start with T3.rnNext is Null
Connect by rnNext = prior rnFirst 得到T4
合成一个语句：
Select T3.CID, T3.AD, sys_connect_by_path(T3.AN ||’(‘ || T3.AS || ‘)’, ‘、’) Area
From (Select T2.CID, T2.AN, T2.AS, T2.AD, T2.rnFirst, lead(T2.rnFirst) over(partition by T2.CID,T2.AD order by T2.rnFirst) rnNext
From (Select T1.CID, T1.AN, T1.AS, T1.AD, row_number() over(order by T1.CID,T1.AD) rnFirst
From T1
) T2
) T3
Start with T3.rnNext is Null
Connect by rnNext = prior rnFirst
方法2：10G 中的wmsys.wm_concat 方法
Select T1.CID, T1.Ad, wmsys.wm_concat(T3.AN ||'('||t1.AS||')') over(partition by T1.CID, T1.AD order by rownum) Area
From T1
构建树的结果也可以保存到一个临时表里。
第四步、分组取最大值：
Select T4.CID, T4.AD, ltrim(max(t4.Arear), ‘、’) Area From T4 得到T5
| CID
|
AD
|
Area
|
| 1
|
1
|
A3(已同步)
|
| 1
|
0
|
A1(同步中)、A2(同步中)、A4(同步中)、A5(未同步)
|
| 2
|
0
|
A1(未同步)、A2(未同步)、A2(未同步)、A4(未同步)、A5(未同步)
|
第四步：行转列
Select T5.CID, max(decode(T5.AD, ‘1’, T5.Area, ‘’)) v1, max(decode(T5.AD, ‘0’, T5.Area, ‘’)) v2
From T5
Group by T5.CID
| CID
|
V1
|
V2
|
| 1
|
A3(已同步)
|
A1(同步中)、A2(同步中)、A4(同步中)、A5(未同步)
|
| 2
|
A1(未同步)、A2(未同步)、A2(未同步)、A4(未同步)、A5(未同步)
|