Question

我有一个非常简单的查询，我用来获得两个单独的计数。我遇到的问题是，通过加入，我得到了不准确的数字偏差。

有没有更好的方法从不同的表中获取两个单独的计数？

SELECT   g.[group] AS groupName,
         COUNT(d.departmentDesc) AS total,
         COUNT(skg.Skillset) AS totalSkillsets
FROM     dbo.TFS_Groups AS g
         LEFT OUTER JOIN
         TFS_Departments_Groups AS d
         ON g.[group] = d.tfsGroup
         LEFT OUTER JOIN
         TFS_Skillsets_Groups AS skg
         ON skg.tfsGroup = g.[Group]
GROUP BY g.[group]
ORDER BY groupName ASC;

Answer 1

如果您使用唯一值后，可以将DISTINCT放入COUNT中即。

SELECT   g.[group] AS groupName,
         COUNT(DISTINCT d.departmentDesc) AS total,
         COUNT(DISTINCT skg.Skillset) AS totalSkillsets
...

我怀疑问题在于你有两个一对多的关系导致额外的行。

减少这种情况的方法是在子查询中预聚合，即

SELECT   g.[group] AS groupName,
         d.vol AS total,
         skg.vol AS totalSkillsets

FROM     dbo.TFS_Groups AS g

         LEFT OUTER JOIN
         (SELECT tfsgroup, count(*) as Vol 
         FROM   TFS_Skillsets_Groups
         group by tfsgroup) AS skg
         ON skg.tfsGroup = g.[Group]

         LEFT OUTER JOIN
         (SELECT tfsgroup, count(*) as Vol 
         FROM   TFS_Departments_Groups 
         group by tfsgroup) AS d
         ON d.tfsGroup = g.[Group]

ORDER BY groupName ASC;

Answer 2

尝试APPLY：

SELECT   g.[group] AS groupName,
         ca1.c1 AS total,
         ca2.c2 AS totalSkillsets
FROM     dbo.TFS_Groups AS g
OUTER APPLY (SELECT COUNT(*) AS c1 FROM TFS_Departments_Groups AS d WHERE g.[group] = d.tfsGroup) ca1
OUTER APPLY (SELECT COUNT(*) AS c2 FROM TFS_Skillsets_Groups AS skg WHERE g.[group] = skg.tfsGroup) ca2

SQL查询使用连接获取计数

2 个答案: