从重复结果中选择具有最大日期的行

Question

注意：我正在使用 SQL 2000。

我想通过为此加入执行类似的操作来选择 t2.JobCompletionTime 的 MAX 日期的行，但我知道这不是正确的代码/语法，因为它不起作用......但我希望它是明确我想要做什么。

LEFT OUTER JOIN (
  SELECT *
  FROM DSM_StandardProcedureActivityView
  WHERE
  ( (PackageName = 'Adobe Acrobat 10'  AND PackageVersion = '-1.0') AND
    (MAX (JobCompletionTime))
  )
) t2 ON t1.UUID = t2.UUID

我现在的完整工作 SQL 代码是：

SELECT 
t1.Business,
t1.[Sub-Business],
t1.Pole,
t1.DomainManager,
t1.ScalabilityServer,
t1.Location,
t1.Country,
t1.Name,
t3.VolumeName,
t3.VolumeFreeInMB,
t3.VolumeSizeInMB,
FROM (
  SELECT *
  FROM DSM_StandardGroupMembersView
  WHERE 
  (  GroupName = 'Adobe Acrobat'
  )
) t1
LEFT OUTER JOIN (
  SELECT *
  FROM DSM_StandardProcedureActivityView
  WHERE
  ( (PackageName = 'Adobe Acrobat 10'  AND PackageVersion = '-1.0')
  )
) t2 ON t1.UUID = t2.UUID
LEFT OUTER JOIN (
  SELECT *
  FROM DSM_StandardHardwareDiskUsageView
  WHERE VolumeName = 'C:\'
) t3 ON t1.Name = t3.Name
ORDER BY t1.Business, t1.[Sub-Business], t1.Pole, t1.DomainManager, t1.ScalabilityServer, t1.Country, t1.Location, [Job Status], t1.Name

Answer 1

逐步完成此操作，以便更明显地了解我可能在此处出现解释错误的地方...

第 1 步：稍微美化您的查询 + 省略 ORDER BY，因为您可以稍后再次轻松添加该查询

SELECT t1.Business,
       t1.[Sub-Business],
       t1.Pole,
       t1.DomainManager,
       t1.ScalabilityServer,
       t1.Location,
       t1.Country,
       t1.Name,
       t3.VolumeName,
       t3.VolumeFreeInMB,
       t3.VolumeSizeInMB

  FROM DSM_StandardGroupMembersView  t1

  LEFT OUTER JOIN DSM_StandardProcedureActivityView t2
               ON t2.PackageName = 'Adobe Acrobat 10'  
              AND t2.PackageVersion = '-1.0'
              AND t2.UUID = t1.UUID

  LEFT OUTER JOIN DSM_StandardHardwareDiskUsageView t3
               ON t3.VolumeName = 'C:'
              AND t3.Name = t1.Name

 WHERE t1.GroupName = 'Adobe Acrobat'

我理解您的问题的方式是您希望将 DSM_StandardProcedureActivityView t2 中的记录限制为仅代表最新情况的记录（基于 t2.JobCompletionTime 字段）。对 ？现在要么我们先把它分组，然后加入那些匹配的（通过 UUID），或者我们反过来想，首先让列表限制为所需的 UUID，然后去掉所有那些 有更新的版本。 由于您使用的是 SQL2000，我建议使用后者，因为我不确定查询优化器是否能够优化前者。 （随意测试=）

然后看起来像这样：

SELECT t1.Business,
       t1.[Sub-Business],
       t1.Pole,
       t1.DomainManager,
       t1.ScalabilityServer,
       t1.Location,
       t1.Country,
       t1.Name,
       t3.VolumeName,
       t3.VolumeFreeInMB,
       t3.VolumeSizeInMB

  FROM DSM_StandardGroupMembersView  t1

  LEFT OUTER JOIN DSM_StandardProcedureActivityView t2
               ON t2.PackageName = 'Adobe Acrobat 10'  
              AND t2.PackageVersion = '-1.0'
              AND t2.UUID = t1.UUID
              -- limit to latest records only!
              AND NOT EXISTS ( SELECT *
                                 FROM DSM_StandardProcedureActivityView t2_newer
                                WHERE t2_newer.UUID = t2.UUID
                                  AND t2_newer.JobCompletionTime > t2.JobCompletionTime )

  LEFT OUTER JOIN DSM_StandardHardwareDiskUsageView t3
               ON t3.VolumeName = 'C:'
              AND t3.Name = t1.Name

 WHERE t1.GroupName = 'Adobe Acrobat'

希望这会有所帮助。

Answer 2

您发布了太多不相关的 SQL，我无法涉足，所以我将修复您的摘录：

LEFT OUTER JOIN (
  SELECT * FROM (
    SELECT *
    FROM DSM_StandardProcedureActivityView
    WHERE PackageName = 'Adobe Acrobat 10'
    AND PackageVersion = '-1.0'
    ORDER BY JobCompletionTime DESC) x
   ) y
   GROUP BY PackageName, PackageVersion
) t2 ON t1.UUID = t2.UUID

这里发生的是内部查询首先获取按最新 JobCompletionTime 排序的行，下一级查询在匹配列上使用 group by，这在 mysql 中为您提供 first 匹配行（最新的，因为我们对行进行了排序）

Answer 3

一个名为 ROW_NUMBER() 的排名函数可以在这里为您提供帮助

LEFT OUTER JOIN (
    SELECT 
        *, 
        rowNumber = ROW_NUMBER() OVER (ORDER BY JobCompletionTime DESC)
    FROM DSM_StandardProcedureActivityView
    WHERE PackageName = 'Adobe Acrobat 10'
    AND PackageVersion = '-1.0'
) t2 ON t1.UUID = t2.UUID AND rowNumber = 1

按 JobCompletionTime 排序（降序）应将最高值作为第 1 行；因此在连接子句中添加了额外的子句。