SQL Query根据列值拆分数据

时间:2014-10-15 14:33:02

标签: sql sql-server excel

我有以下数据:

SR_ID   OPEN_DATE   OPEN_WEEK   PRIORITY    TTRespond
------- ----------- ----------- ----------- -------------
24720   01/10/2014  40          P2 - High   3.867066667
24437   11/09/2014  37          P2 - High   418.1992333
24007   04/08/2014  32          P1 - Urgent 12571.28308
24628   25/09/2014  39          P2 - High   3.0407
24446   12/09/2014  37          P2 - High   2694.122933
24420   10/09/2014  37          P3 - Normal 
24479   15/09/2014  38          P2 - High   90.56748333
24924   15/10/2014  42          P3 - Normal 26.51546667
24706   01/10/2014  40          P2 - High   

我正在尝试编写一个查询,该查询将根据OPEN_WEEK和PRIORITY中的值在单独的列中返回最后一列(TTRespond)中的数据。

我想像这样返回数据:

Priority    P1 - Urgent P1 - Urgent P1 - Urgent P1 - Urgent P1 - Urgent P1 - Urgent
----------- ----------- ----------- ----------- ----------- ----------- -----------
Week #      31          32          33          34          35          36

我不想总结数据(即Count Average等),我只想返回符合这两个条件的所有行。我在Excel中使用基于数组的公式获得了结果,但它非常慢并且无法处理源记录集(仅25,000行)。

非常感谢任何帮助。

由于

1 个答案:

答案 0 :(得分:1)

你并不完全清楚你在问什么。但是假设您正在寻找类似这样的输出,其中优先级/周数是列,并且所有可能的匹配值都列在行中:

P2 - High, Week 37 | P2 - High, Week 38 | P2 - High, Week 39 | P2 - High, Week 40
---------------------------------------------------------------------------------
418.1992333        | 90.56748333        | 3.0407             | 3.867066667
2694.122933        |                    |                    |

这种"行/列反转"被称为数据透视表。通过在SQL Server中以下列方式巧妙地使用group bymax子句,可以完成简单的数据透视表:

;with T as (
    select
        row_number() over (partition by priority, open_week order by open_date) ix
        ,T.*
    from MyTable T
    where TTrespond is not null
), R as (
    select distinct ix from T
)
select
    R.ix
    ,max(case when [OPEN_WEEK]=37 and T.ix=R.ix and [PRIORITY]='P2 - High' then TTRespond end) as [P2 - High, Week 37]
    ,max(case when [OPEN_WEEK]=38 and T.ix=R.ix and [PRIORITY]='P2 - High' then TTRespond end) as [P2 - High, Week 38]
    ,max(case when [OPEN_WEEK]=39 and T.ix=R.ix and [PRIORITY]='P2 - High' then TTRespond end) as [P2 - High, Week 39]
    ,max(case when [OPEN_WEEK]=40 and T.ix=R.ix and [PRIORITY]='P2 - High' then TTRespond end) as [P2 - High, Week 40]
from T, R
group by R.ix

可以通过添加在上述示例之后建模的更多列组合来扩展可能的值。