将 SQL 表与其自身进行比较以查找缺失的时间表日期答案

【问题标题】：Compare SQL table to itself to find missing timesheet dates将 SQL 表与其自身进行比较以查找缺失的时间表日期
【发布时间】：2017-04-05 21:34:39
【问题描述】：

下面的 SQL 查询有效，但由于使用了子查询，它需要很长时间才能响应。我希望有人能为我找到优化此查询的方法。

SELECT     a.Userid, a.Date, a.[Total Hrs], a.[Load Date],
                  c.[First Name], c.[Last Name] 
FROM         dbo.TimeSheetReport AS a 
                  INNER JOIN
                  Candidates AS c ON a.Userid = c.Userid 
AND a.Date >= c.[Confirmed Start Date] 
AND (a.Date <= c.[Planned End Date] or c.[Planned End Date] is null)
WHERE a.[Load Date] = @ddlendDate 
AND a.Date not in 
(select b.Date from dbo.TimeSheetReport AS b 
where b.[Load Date] = @ddlstartDate and a.Userid = b.Userid)
AND a.Date < (select MAX(date) from dbo.TimeSheetReport AS c 
where c.[Load Date] = @ddlstartDate)

每个数据子集都是根据数据的加载时间（即加载日期）创建的。它从当前加载 (@ddlendDate) 中查找之前加载 (@ddlstartDate) 中缺少的任何日期。

我尝试在其自身上使用 LEFT OUTER JOIN，但返回的记录为零，我知道它们存在。

有什么想法吗？

【问题讨论】：

标签： sql sql-server-2008 stored-procedures

【解决方案1】：

您可能希望对第一个子查询使用 公用表表达式（有关详细信息，请参阅此链接：https://technet.microsoft.com/en-us/library/ms190766(v=sql.105).aspx）或临时表。

第二个可以存储在变量中，不需要定义为子查询（所以只会执行一次）

declare @MaDate DateTime 
select @MaDate = MAX([date])  from dbo.TimeSheetReport AS c  where c.[Load Date] = @ddlstartDate

SELECT     a.Userid, a.Date, a.[Total Hrs], a.[Load Date],
                  c.[First Name], c.[Last Name] 
FROM         dbo.TimeSheetReport AS a 
 INNER JOIN   Candidates AS c ON a.Userid = c.Userid 
        AND a.Date >= c.[Confirmed Start Date] 
        AND (a.Date <= c.[Planned End Date] or c.[Planned End Date] is null)
WHERE a.[Load Date] = @ddlendDate 
    AND a.Date not in 
        (select b.Date from dbo.TimeSheetReport AS b 
         where b.[Load Date] = @ddlstartDate and a.Userid = b.Userid)
    AND a.Date < @MaDate

【讨论】：

感谢部分解决方案。你的第一个建议不起作用，这个子查询是最需要解决的。第二个看起来很有帮助。

【解决方案2】：

将 NOT IN 转换为 NOT EXISTS 将查询执行时间从 12 多分钟减少到 10 秒以下。

SELECT     a.Userid, a.Date, a.[Total Hrs], a.[Load Date],
                  c.[First Name], c.[Last Name] 
FROM         dbo.TimeSheetReport AS a 
                  INNER JOIN
                  Candidates AS c ON a.Userid = c.Userid 
AND a.Date >= c.[Confirmed Start Date] 
AND (a.Date <= c.[Planned End Date] or c.[Planned End Date] is null)
WHERE a.[Load Date] = @ddlendDate 
AND NOT EXISTS (select 1 from dbo.TimeSheetReport AS b 
WHERE b.[Load Date] = @ddlstartDate AND a.Userid = b.Userid AND a.Date = b.Date)
AND a.Date < (select MAX(date) from dbo.TimeSheetReport AS c 
WHERE c.[Load Date] = @ddlstartDate)

【讨论】：