将一些行值透视到列

时间:2020-05-27 06:17:33

标签: sql sql-server ssis-2012

我正在尝试将一些行值转换为列,这是我要实现的目标。

我当前的模式:

+------+----------+
|  ID  |   name   |
+------+----------+
|   01 |  Vsp lan |
| 0121 |  abn     |
| 0122 |  abb     |
| 0123 |  vsp     |
|   02 |  Apn lan |
| 0211 |  add     |
| 0221 |  acd     |
+------+----------+

这是我要实现的目标:

+-----+--------+-------+---------+
| kod |   ID   | name  |   lan   |
+-----+--------+-------+---------+
|  01 |   0121 |   abn | vsp lan |
|  01 |   0122 |   abb | vsp lan |
|  01 |   0123 |   vsp | vsp lan |
|  02 |   0211 |   add | Apn lan |
|  02 |   0221 |   acd | Apn lan |
+-----+--------+-------+---------+

但是当name和lan具有相似的值时,它将跳过行,在这种情况下,它将跳过名称值为vsp的行。

DECLARE @table TABLE (ID VARCHAR(5),[name] VARCHAR(10));
INSERT INTO @table
VALUES
 ('01','Vsp Ian')
,('0121','abn')
,('0122','abb')
,('0123','vsp')

,('02','Apn Ian')
,('0211','add')
,('0221','acd')
;

SELECT a.id as kod, b.id as ID, B.name as name, a.name as lan
FROM @table a
inner join @table b on CHARINDEX(a.id,b.id) = 1 and CHARINDEX(b.name,a.name) = 0

5 个答案:

答案 0 :(得分:1)

从您的示例中,假设van_ian kod是名称ID的开头,则假定您要使用其他名称映射Van_ian。例如01是0121、0122的开头。

DECLARE @table table(id CHAR(10), name varchar(10))

INSERT INTO @table
values
('01','vsp ian'),
('0121','abn'),
('0122','abb'),
('02','vsp ian'),
('0211','add'),
('0221','acd');

SELECT ct.id as kod,ot.id, ot.name,ct.name as ian FROM @table as ot
CROSS JOIN
(SELECT distinct id, name from @table WHERE name like 'vsp%'
) as ct(id,name)
WHERE ot.name not like 'vsp%'
and LEFT(ot.id,2) = ct.id
+-----+--------+-------+---------+
| kod |   id   | name  |   ian   |
+-----+--------+-------+---------+
|  01 |   0121 |   abn | vsp ian |
|  01 |   0122 |   abb | vsp ian |
|  02 |   0211 |   add | vsp ian |
|  02 |   0221 |   acd | vsp ian |
+-----+--------+-------+---------+

答案 1 :(得分:1)

这解决了我的问题。

DECLARE @table TABLE (ID VARCHAR(5),[name] VARCHAR(10));
INSERT INTO @table
VALUES
 ('01','Vsp Ian')
,('0121','abn')
,('0122','abb')
,('0123','vsp')
,('02','Apn Ian')
,('0211','add')
,('0221','acd')
,('03','Ubb Ian')
,('0301','afg')
,('0302','ampx');


SELECT
 kod = ct.id
,ot.ID
,ot.[name]
,Ian = ct.[name]
FROM @table ot
CROSS JOIN (SELECT ID,[name] FROM @table WHERE [name] LIKE '%[lan]%') ct
WHERE len(ot.ID) > 2 
AND LEFT(ot.id,2) = ct.id;

答案 2 :(得分:0)

我能想到的简单方法是自我连接,其中一个id起始于另一个

SELECT a.id as kod, b.id as ID, B.name as name, a.name as laen
FROM table_name a
inner join table_name b on CHARINDEX(a.name,b.name) = 1 and CHARINDEX(b.name,a.name) = 0

这应该适用于您的示例,但是如果ID也较长,则结果可能会很奇怪

答案 3 :(得分:0)

我正在观察ID值中的模式。例如,可以将ID 0121, 0122归类为ID 01。类似地,可以将ID 0211, 0221归类为ID 02。 基于此假设,此查询应执行以下操作:

select 
    tb1.ID as kod, 
    tb2.ID as ID, 
    tb2.name as name, 
    tb1.name as laen 
from tableName tb1 join tableName tb2 
on tb1.ID != tb2.ID and LOCATE(tb1.ID,tb2.ID) = 1

答案 4 :(得分:0)

Venkataraman R,您的回答非常接近,我对其进行了一些调整以使其与他的输出保持一致。

DECLARE @table TABLE (ID VARCHAR(5),[name] VARCHAR(10));
INSERT INTO @table
VALUES
 ('01','Vsp_Ian')
,('0121','abn')
,('0122','abb')
,('02','Apn_Ian')
,('0211','add')
,('0221','acd')
,('03','Ubb_Ian')
,('0301','afg')
,('0302','ampx');

SELECT
 kod = ct.id
,ot.ID
,ot.[name]
,Ian = ct.[name]
FROM @table ot
CROSS JOIN (SELECT ID,[name] FROM @table WHERE [name] LIKE '%[_]%') ct
WHERE ot.[name] NOT LIKE '%[_]%'
AND LEFT(ot.id,2) = ct.id;