SQL选择复杂逻辑

我有一个复杂的场景，我会尽力解释

我在ProjectPhases表中有一个名为“Phase”的列，其中包含关于项目阶段的描述。然后，有一个名为DailyReport的表包含诸如描述、QTY1、QTY2、PHASEID（ProjectPhases表中的外键）等字段

现在，前端的用户将为一周中的每一天创建每日报告记录。现在，我编写查询来总结包含ProjectPhases数据和daily report数据的日报

       select P.projectname ,PP.Description as Phase,DD.description, DD.QTYSunday,DD.QTYSMonday,DD.QTYTues,
       from 
       Document_DailyReport DD  LEFT OUTER JOIN
       ProjectPhases PP on PP.Id=DD.PhaseId LEFT OUTER JOIN
       Projects P  on P.Id=DD.ProjectId

查询的输出采用以下格式

  ProjectName Phase Description QTYSunday QTYMonday QtyTues  
  Project1   Phase 1 Qty-SUNDAy 10                      
  Project1   Phase 1 Qty-Monday              10          
  Project1   Phase 1 Qty-Monday                        10

    ProjectName Phase  QTYSunday QTYMonday QtyTues  
    Project1   Phase 1    10       10        10   

现在我希望输出的格式如下

  ProjectName Phase Description QTYSunday QTYMonday QtyTues  
  Project1   Phase 1 Qty-SUNDAy 10                      
  Project1   Phase 1 Qty-Monday              10          
  Project1   Phase 1 Qty-Monday                        10

    ProjectName Phase  QTYSunday QTYMonday QtyTues  
    Project1   Phase 1    10       10        10   

我想把某一特定阶段的每日报告的所有记录放在一行中，就像上面所说的那样

谢谢

编辑：

我的完整问题如下

SELECT P.projectname ,
       PP.Description AS Phase,
       DD.Description,
       DD.DocNumber,

  (SELECT CompanyName
   FROM Companies
   WHERE Id=dbo.GetCompanyIdByUser(DD.InsertedBy)) AS CreatedBy,
       DD.ReportDate,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, -(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantitySun,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 1-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantityMon,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 2-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantityTues,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 3-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantityWed,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 4-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantityThur,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 5-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantityFri,

  (SELECT sum(quantity)+sum(cast(field9 AS INT))
   FROM Document_DailyReportOnSite
   WHERE DailyReportId=DD.Id
     AND ClassificationId=1
     AND DATEDIFF(DAY,reportdate,DATEADD(dd, 6-(DATEPART(dw, (@Date)-1)), @Date)) =0) AS StaffQuantitySat ,
FROM Document_DailyReport DD
LEFT OUTER JOIN ProjectPhases PP ON PP.Id=DD.PhaseId
LEFT OUTER JOIN Projects P ON P.Id=DD.ProjectId
WHERE DD.ReportDate BETWEEN DATEADD(dd, -(DATEPART(dw,@Date)-1), @Date) AND DATEADD(dd, 6-(DATEPART(dw, @Date)-1), @Date)

输出

 projectname                Phase                                             description                     CreatedBy  0StaffQuantityMon StaffQuantityTues  StaffQuantityWed  StaffQuantityThur
   Bollywood Park   (MCC-ARCO) Bollywood Theatre Main Contractor Package    Daily Report as on 16-Jun-2014  ARCO Contracting    22     NULL     NULL    NULL
   Bollywood Park   (MCC-ARCO) Bollywood Theatre Main Contractor Package    Daily Report as om 17-Jun-2014  ARCO Contracting    NULL        23       NULl        NULL
  Bollywood Park    (MCC-ARCO) Bollywood Theatre Main Contractor Package    Daily Report as on 18.06.2014   ARCO Contracting    NULL       NULL      NULL    23

---

使用

分组依据

功能

试试这个：

SELECT P.projectname, PP.Description AS Phase, 
       SUM(DD.QTYSunday) AS QTYSunday, 
       SUM(DD.QTYSMonday) AS QTYSMonday, 
       SUM(DD.QTYTues) AS QTYTues
FROM Projects P 
INNER JOIN Document_DailyReport DD ON P.Id = DD.ProjectId
LEFT OUTER JOIN ProjectPhases PP ON PP.Id = DD.PhaseId 
GROUP BY P.projectname, PP.Description;

---

我可以看到你的描述包括列名，如“数量星期日”

---

您可以使用描述字段上的轴来返回所需的结果集。有大量文档可供您使用。

您可以尝试以下方法：

SELECT ProjectName, Phase, SUM(StaffQuantitySun) AS StaffQuantitySun, SUM(StaffQuantityMon) AS StaffQuantityMon, SUM(StaffQuantityTues) AS StaffQuantityTues, SUM(StaffQuantityWed) AS StaffQuantityWed,
SUM(StaffQuantityThur) AS StaffQuantityThur, SUM(StaffQuantityFri) AS StaffQuantityFri, SUM(StaffQuantitySat) AS StaffQuantitySat
(
select P.projectname ,PP.Description as Phase,DD.Description,DD.DocNumber,
(Select CompanyName from Companies where Id=dbo.GetCompanyIdByUser(DD.InsertedBy)) As CreatedBy,DD.ReportDate,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, -(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantitySun,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 1-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantityMon,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 2-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantityTues,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 3-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantityWed,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 4-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantityThur,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 5-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantityFri,

(select sum(quantity)+sum(cast(field9 as INT)) from Document_DailyReportOnSite  where DailyReportId=DD.Id  and ClassificationId=1 and DATEDIFF(day,reportdate,DATEADD(dd, 6-(DATEPART(dw, (@Date)-1)), @Date)) =0) as StaffQuantitySat ,

from 
    Document_DailyReport DD  LEFT OUTER JOIN
    ProjectPhases PP on PP.Id=DD.PhaseId LEFT OUTER JOIN
    Projects P  on P.Id=DD.ProjectId
Where DD.ReportDate between DATEADD(dd, -(DATEPART(dw,@Date)-1), @Date) and DATEADD(dd, 6-(DATEPART(dw, @Date)-1), @Date)
) DerivedTable
GROUP BY ProjectName, Phase

---

基本上，我已经将您的查询嵌入到一个派生表中，然后按项目名称和阶段进行分组，然后将您希望的所有列汇总到一行中。

我认为这不会产生预期的结果，我可能是错的，尽管QTYSunday，QTYMonday都是子查询，我把它们作为直接字段，以便使我的问题不那么复杂。由于它们是子查询，我无法对它们应用聚合函数。这称为

PIVOT

查询，SQL Server实际上有一些函数可以帮助实现这一点。可能的重复我不确定是否可以使用PIVOT完成。我的输出中的值不是常量，无法将其转换为PIVOT，这只是一个示例。描述将不是固定不变的。请在我的问题的末尾找到原始输出。很酷，不用担心，伙计。注意，考虑到子查询中的日期数学，我希望这会像糖浆一样缓慢地覆盖任何实际大小的数据集。根据实际日期进行分组（可能是使用

PIVOT

的一部分）可能会产生更好的结果。只是不要包含任何不属于最终输出的列。