我正在尝试根据3个值透视一组数据,以便每一行代表与每个ID相关的所有个人资料网址,而每一列都与个人资料代码相关
这个问题与我要解决的问题非常相似,但是它不是根据另一个值Pivot values on column based on grouped columns in SQL
对正在旋转的列进行分组因此,给出了以下示例表:
Id ProfileCode ProfileURL
-------------------------------------------------------
7ADC7368 IA http://www.domainIA.com/profile1
5C627D6F IA http://www.domainIA.com/profile2
5C627D6F AG http://www.domainAG.com/profile1
5C627D6F AF http://www.domainAF.com/profile1
664B4AE9 IA http://www.domainIA.com/profile3
664B4AE9 AF http://www.domainAF.com/profile2
我希望将其转换为下表:
Id IA AG AF
-------------------------------------------------------------------------------------------------------------
7ADC7368 http://www.domainIA.com/profile1 null null
5C627D6F http://www.domainIA.com/profile2 http://www.domainAG.com/profile1 http://www.domainAF.com/profile1
664B4AE9 http://www.domainIA.com/profile3 null http://www.domainAF.com/profile2
这是我一直在尝试的代码,但是我找不到将枢轴与配置文件URL及其关联的配置文件代码之间的关联相关联的方法。
declare @tmp TABLE (Id NVARCHAR(15), ProfileCode NVARCHAR(2), ProfileURL NVARCHAR(50))
insert into @tmp (Id, ProfileCode, ProfileURL)
values ('7ADC7368', 'IA', 'http://www.domainIA.com/profile1'),
('5C627D6F', 'IA', 'http://www.domainIA.com/profile2'),
('5C627D6F', 'AG', 'http://www.domainAG.com/profile1'),
('5C627D6F', 'AF', 'http://www.domainAF.com/profile1'),
('664B4AE9', 'IA', 'http://www.domainIA.com/profile3'),
('664B4AE9', 'AF', 'http://www.domainAF.com/profile2')
select
pvt.id,
CASE
WHEN ProfileCode = 'IA' THEN ProfileURL
END AS 'IA',
CASE
WHEN ProfileCode = 'AF' THEN ProfileURL
END AS 'AF',
CASE
WHEN ProfileCode = 'AG' THEN ProfileURL
END AS 'AG'
from (
select
Id, ProfileCode, ProfileURL
,ROW_NUMBER() over(partition by ProfileCode order by ProfileURL) as RowNum
from
@tmp
) a
pivot (MAX(ProfileCode) for RowNum in ('IA', 'AF', 'AG') as pvt
如果能为我提供帮助或帮助,我将不胜感激。
答案 0 :(得分:1)
只需使用条件聚合:
SELECT id,
MAX(CASE WHEN ProfileCode = 'IA' THEN ProfileURL END) AS IA,
MAX(CASE WHEN ProfileCode = 'AF' THEN ProfileURL END) AS AF,
MAX(CASE WHEN ProfileCode = 'AG' THEN ProfileURL END) AS AG
FROM @tmp t
GROUP BY id;
对于给定ID ,如果您有多个相同的代码,则只需要ROW_NUMBER()
,并且您希望将结果放在单独的行中。您的样本数据和当前逻辑表明情况并非如此。