具有相似列名的动态数据透视表

Question

我正在寻找一种将不同数量的行转换为 sql server 2008 R2 中的列的方法。我在查询中创建了数据列 RANK，因为最终我希望将透视列名称标记为 RANK 列中的值。然后，如果我能以某种方式 STUFF 将其他 3 个字段值一起合并到一个字段中，我将能够遍历后端语言中的行并适当地拆分字段结果。

这是当前的数据集：

我希望枢轴的最终结果产生这样的结果：

我还没有发现任何关于能够以这种 "dynamic" 方式旋转的信息。任何帮助将不胜感激。

Answer 1

正如我上面提到的，您需要区分每个 Rank 值。您已经说过这是一个计算值，您可以在每个值的末尾添加一个数字。添加该数字后，您仍然需要对其进行旋转。

首先看到这一点的最简单方法是先编写查询的硬编码版本。

示例数据：

create table yourdata
(
    id int,
    code varchar(50),
    created datetime,
    [rank] varchar(50)
);

insert into yourdata
select 285856, 'J7609', '2015-01-19', 'Principle' union all
select 285856, 'J7613', '2015-01-19', 'Other' union all
select 285856, 'J0456', '2015-01-19', 'Other' union all
select 285856, 'J0694', '2015-01-19', 'Other' union all
select 285856, 'J1885', '2015-01-19', 'Other' union all
select 285856, 'J2060', '2015-01-19', 'Other' union all
select 285856, 'J2930', '2015-01-19', 'Other';

静态查询：

select Principle_1, Other_1, 
    Other_2, Other_3, Other_4,
    Other_5, Other_6
from
(
    -- using row_number to get unique id for each rank
    select 
        data = cast(id as varchar(10)) +' | '+ code +' | '+ convert(varchar(10), created, 112),     
        [rank] = [rank] + '_' +cast(row_number() over(partition by id, [rank] 
                                                        order by id) as varchar(10))
    from yourdata
) d
pivot
(
    max(data)
    for [rank] in (Principle_1, Other_1, Other_2, Other_3, Other_4,
                    Other_5, Other_6)
) p;

现在要执行此动态操作，您将使用列名创建一个 sql 字符串，然后执行该字符串：

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT  ',' + QUOTENAME([rank] + '_' +cast(rn as varchar(10))) 
                    from
                    (
                        select [rank],
                            rn = row_number() over(partition by id, [rank] 
                                                    order by id) 
                        from  yourdata
                    ) d
                    group by [rank], rn
                    order by rn, [rank] desc
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query = 'SELECT ' + @cols + '
            from 
             (
                select 
                    data = cast(id as varchar(10)) +'' | ''+ code +'' | ''+ convert(varchar(10), created, 112),     
                    [rank] = [rank] + ''_'' +cast(row_number() over(partition by id, [rank] 
                                                                    order by id) as varchar(10))
                from yourdata
            ) x
            pivot 
            (
                max(data)
                for [rank] in (' + @cols + ')
            ) p '

exec sp_executesql @query;

这会得到一个结果：

+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+
|        Principle_1        |          Other_1          |          Other_2          |          Other_3          |          Other_4          |          Other_5          |          Other_6          |
+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+
| 285856 | J7609 | 20150119 | 285856 | J7613 | 20150119 | 285856 | J0456 | 20150119 | 285856 | J0694 | 20150119 | 285856 | J1885 | 20150119 | 285856 | J2060 | 20150119 | 285856 | J2930 | 20150119 |
+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+---------------------------+

具有相似列名的动态数据透视表

Dynamic pivot with similar column names

sql

sql-server

pivot

sql-server-2008-r2