Sql 最大日期和在单独列上连接

有人能帮我解决这个问题吗？我需要用我的数据做两件事

我需要得到最长的日期

我需要连接Proj_Use列中的值

我的数据如下所示：

Report_Date  |  Program  |  Project  |  Proj_Mgr  |  Region  |  Proj_Use

6/5/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU1

6/4/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU2

6/3/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU3

6/5/2017  |  PG2  |  Prj02  |  Monica  |  West  | PU4

6/4/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU5

6/3/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU6

Report_Date Program Project Proj_MgrRegion  Proj_Use

6/5/2017 | PG1 | Prj01 | Anne | East | PU1, PU2, PU3

6/5/2017 | PG2 | Prj02 | Monica | West | PU4, PU5, PU6

SELECT pt.*,
       STUFF((SELECT ',' + pt2.[Proj_Use]
              FROM project_table pt2
              WHERE pt2.Program = pt.Program
              ORDER BY pt2.[Report Date]
              FOR XML PATH ('')
             ), 1, 1, '') AS [Projects Used]
FROM (SELECT pt.*,
             ROW_NUMBER() OVER (PARTITION BY Program ORDER BY [Report Date] DESC) as seqnum
      FROM project_table t
     ) pt
WHERE seqnum = 1;

我的结果应该如下所示：

Report_Date  |  Program  |  Project  |  Proj_Mgr  |  Region  |  Proj_Use

6/5/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU1

6/4/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU2

6/3/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU3

6/5/2017  |  PG2  |  Prj02  |  Monica  |  West  | PU4

6/4/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU5

6/3/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU6

Report_Date Program Project Proj_MgrRegion  Proj_Use

6/5/2017 | PG1 | Prj01 | Anne | East | PU1, PU2, PU3

6/5/2017 | PG2 | Prj02 | Monica | West | PU4, PU5, PU6

SELECT pt.*,
       STUFF((SELECT ',' + pt2.[Proj_Use]
              FROM project_table pt2
              WHERE pt2.Program = pt.Program
              ORDER BY pt2.[Report Date]
              FOR XML PATH ('')
             ), 1, 1, '') AS [Projects Used]
FROM (SELECT pt.*,
             ROW_NUMBER() OVER (PARTITION BY Program ORDER BY [Report Date] DESC) as seqnum
      FROM project_table t
     ) pt
WHERE seqnum = 1;

我可以使用以下代码获取最大日期：

    SELECT   t.[Report_Date]
        ,t.[Program]
        ,t.[Project]
        ,t.[Proj_Mgr]
        ,t.[Region]
        ,t.[Proj_Use]
    FROM project_table t
    JOIN
    (SELECT MAX([Report_Date]) as [Report Date], [Program]
        FROM project_table
        GROUP BY [Program]) max
        ON t.[Report_Date] = max.[Report Date]
        AND t.[Program] = max.[Program]

我在论坛上搜索了一种连接方法，找到了sql server的group_concat方法。我尝试将其添加到代码中，并得出以下结论：

    SELECT   t.[Report_Date]
        ,t.[Program]
        ,t.[Project]
        ,t.[Proj_Mgr]
        ,t.[Region]
        ,STUFF ((
            SELECT ',' + t.[Proj_Use]
                FROM project_table
                FOR XML PATH ('')), 1,1, '') AS [Project Use]
    FROM project_table t
    JOIN
    (SELECT MAX([Report_Date]) as [Report Date], [Program]
        FROM project_table
        GROUP BY [Program]) max
        ON t.[Report_Date] = max.[Report Date]
        AND t.[Program] = max.[Program]

当我尝试使用真实数据运行此查询时，查询运行了20多分钟而没有生成结果

有没有人能告诉我，我是否正确地构造了这个查询，或者是否有更好的方法来获得我想要的结果

---

谢谢你的帮助

我期待着这样的事情：

Report_Date  |  Program  |  Project  |  Proj_Mgr  |  Region  |  Proj_Use

6/5/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU1

6/4/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU2

6/3/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU3

6/5/2017  |  PG2  |  Prj02  |  Monica  |  West  | PU4

6/4/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU5

6/3/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU6

Report_Date Program Project Proj_MgrRegion  Proj_Use

6/5/2017 | PG1 | Prj01 | Anne | East | PU1, PU2, PU3

6/5/2017 | PG2 | Prj02 | Monica | West | PU4, PU5, PU6

SELECT pt.*,
       STUFF((SELECT ',' + pt2.[Proj_Use]
              FROM project_table pt2
              WHERE pt2.Program = pt.Program
              ORDER BY pt2.[Report Date]
              FOR XML PATH ('')
             ), 1, 1, '') AS [Projects Used]
FROM (SELECT pt.*,
             ROW_NUMBER() OVER (PARTITION BY Program ORDER BY [Report Date] DESC) as seqnum
      FROM project_table t
     ) pt
WHERE seqnum = 1;

---

至少，进行连接的子查询没有到外部查询的关联子句（因此中间字符串可能很长）。可能还有其他问题。

我期待着这样的事情：

Report_Date  |  Program  |  Project  |  Proj_Mgr  |  Region  |  Proj_Use

6/5/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU1

6/4/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU2

6/3/2017  |  PG1  |  Prj01  |  Anne  |  East  |  PU3

6/5/2017  |  PG2  |  Prj02  |  Monica  |  West  | PU4

6/4/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU5

6/3/2017  |  PG2  |  Prj02  |  Monica  |  West  |  PU6

Report_Date Program Project Proj_MgrRegion  Proj_Use

6/5/2017 | PG1 | Prj01 | Anne | East | PU1, PU2, PU3

6/5/2017 | PG2 | Prj02 | Monica | West | PU4, PU5, PU6

SELECT pt.*,
       STUFF((SELECT ',' + pt2.[Proj_Use]
              FROM project_table pt2
              WHERE pt2.Program = pt.Program
              ORDER BY pt2.[Report Date]
              FOR XML PATH ('')
             ), 1, 1, '') AS [Projects Used]
FROM (SELECT pt.*,
             ROW_NUMBER() OVER (PARTITION BY Program ORDER BY [Report Date] DESC) as seqnum
      FROM project_table t
     ) pt
WHERE seqnum = 1;

---

至少，进行连接的子查询没有到外部查询的关联子句（因此中间字符串可能很长）。可能还有其他问题。

您缺少的一件大事是子查询中的

where

子句

使用

试试这个：

select
    ReportDate = max(t.Report_Date)
  , t.Program
  , t.Project
  , t.Proj_Mgr
  , t.Region
  , [Project_Use] = stuff((
   select ',' + i.Proj_Use
   from project_table i
   where i.Program = t.Program
     and i.Project = t.Project
     and i.Proj_Mgr = t.Proj_Mgr
     and i.Region = t.Region
   for xml path(''), type).value('.','nvarchar(max)')
   , 1, 1, '') 
from project_table t
group by t.Program, t.Project, t.Proj_Mgr, t.Region

rextester演示：

返回：

+------------+---------+---------+----------+--------+-------------+
| ReportDate | Program | Project | Proj_Mgr | Region | Project_Use |
+------------+---------+---------+----------+--------+-------------+
| 2017-06-04 | pg1     | Prj01   | Anne     | East   | pu1,pu2,pu3 |
| 2017-06-04 | pg2     | Prj02   | Monica   | West   | pu4,pu5,pu6 |
+------------+---------+---------+----------+--------+-------------+

---

您缺少的一件大事是子查询中的

where

子句

使用

试试这个：

select
    ReportDate = max(t.Report_Date)
  , t.Program
  , t.Project
  , t.Proj_Mgr
  , t.Region
  , [Project_Use] = stuff((
   select ',' + i.Proj_Use
   from project_table i
   where i.Program = t.Program
     and i.Project = t.Project
     and i.Proj_Mgr = t.Proj_Mgr
     and i.Region = t.Region
   for xml path(''), type).value('.','nvarchar(max)')
   , 1, 1, '') 
from project_table t
group by t.Program, t.Project, t.Proj_Mgr, t.Region

rextester演示：

返回：

+------------+---------+---------+----------+--------+-------------+
| ReportDate | Program | Project | Proj_Mgr | Region | Project_Use |
+------------+---------+---------+----------+--------+-------------+
| 2017-06-04 | pg1     | Prj01   | Anne     | East   | pu1,pu2,pu3 |
| 2017-06-04 | pg2     | Prj02   | Monica   | West   | pu4,pu5,pu6 |
+------------+---------+---------+----------+--------+-------------+

---

您正在使用哪个数据库？sql server 2014您正在使用哪个数据库？sql server 2014这对我很有用。我确实想向任何研究该解决方案的人指出，您在for XML语句中包含了一个额外的结束语。要使其正常工作，它应该如下所示：对于xml路径（“”），键入.value（‘.’，'nvarchar（max）’），1，1，“”），也感谢您提供了指向其他信息的链接。非常有帮助。@matt哎呀！至少在演示中是正确的。修正了答案中的拼写错误。这对我很有效。我确实想向任何研究该解决方案的人指出，您在for XML语句中包含了一个额外的结束语。要使其正常工作，它应该如下所示：对于xml路径（“”），键入.value（‘.’，'nvarchar（max）’），1，1，“”），也感谢您提供了指向其他信息的链接。非常有帮助。@matt哎呀！至少在演示中是正确的。修正了答案中的错误。