Question

我试图拉回加载到表中的某些记录的前五个实例。重要的是要注意我试图从同一个表中得到我的结果，我认为有一个问题，我无法弄清楚这是一个表的事实。以下是我的查询示例：

Select * From (
Select Top 5 JobID From Jobs Where JobTypeID = 1 Order By JobID DESC
UNION ALL
Select Top 5 JobID From Jobs Where JobTypeID = 2 Order By JobID DESC
UNION ALL
Select Top 5 JobID From Jobs Where JobTypeID = 3 Order By JobID DESC
UNION ALL
Select Top 5 JobID From Jobs Where JobTypeID = 4 Order By JobID DESC
UNION ALL
Select Top 5 JobID From Jobs Where JobTypeID = 5 Order By JobID DESC
UNION ALL
Select Top 5 JobID From Jobs Where JobTypeID = 6 Order By JobID DESC
) As UnionTable

当我运行这是SQL Server时，我只知道应该有30条记录。我怎样才能确保撤回所有正确的记录？我是否需要在每个子查询中使用Group By子句？

Answer 1

更优雅..

;WITH cte AS
(
    SELECT
         JobID, 
         ROW_NUMBER() OVER (PARTITION BY JobTypeID ORDER BY JobID /* eh? */ DESC) AS rank
    FROM
         Jobs WHERE JobTypeID BETWEEN 1 AND 6
)
SELECT * FROM CTE WHERE rank <= 5

~~为什么你有JobID的WHERE和ORDER？这使得ORDER BY子句毫无意义......~~

Answer 2

分别运行每个SELECT语句，看看每个

是否有5行

运行此项以查看每个的计数，返回的第二列将是JobID

Select COUNT(*),1 From Jobs Where JobID = 1 
UNION ALL
Select COUNT(*),2 From Jobs Where JobID = 2 
UNION ALL
Select COUNT(*),3 From Jobs Where JobID = 3 
UNION ALL
Select COUNT(*),4 From Jobs Where JobID = 4 
UNION ALL
Select COUNT(*),5 From Jobs Where JobID = 5 
UNION ALL
Select COUNT(*),6 From Jobs Where JobID = 6

您也可以像这样运行

Select COUNT(*),JobID
 From Jobs 
 Where JobID between 1 and 6 
 GROUP By JobID

使用UNION ALL和ORDER BY选择Top（某个数字）

2 个答案: