如何修复where子句中的聚合?
Posted
技术标签:
【中文标题】如何修复where子句中的聚合?【英文标题】:How to fix aggregate in where clause? 【发布时间】:2017-04-19 22:16:12 【问题描述】:消息 147,级别 15,状态 1,过程 vw_OverBudget,第 10 行 [批处理开始第 59 行] 聚合可能不会出现在 WHERE 子句中,除非它位于 HAVING 子句或选择列表中包含的子查询中,并且被聚合的列是外部引用。
代码:
CREATE VIEW vw_OverBudget
AS
SELECT
p.projectName,
SUM(a.costtodate) AS sumActivityCosts,
p.fundedbudget
FROM
Project AS p
FULL OUTER JOIN
Activity a ON p.projectid = a.projectid
WHERE
a.activityId IS NULL
AND p.projectid IS NOT NULL
AND SUM(a.costtodate) > p.fundedbudget
GROUP BY
p.projectID
【问题讨论】:
【参考方案1】:此查询不需要full outer join
。根据您的逻辑,left join
似乎是正确的:
Select p.projectName, sum(a.costtodate) AS sumActivityCosts,
p.fundedbudget
From Project p left join
Activity a
on p.projectid = a.projectid
where a.activityId is null and p.projectid is not null
Group By p.projectName, p.fundedbudget
having sum(a.costtodate) > p.fundedbudget;
但是,这对我来说没有意义。如果a.activityId
为空,那么最可能的原因是没有匹配。所以,我很确定你只想要一个没有 where
子句的 inner join
:
Select p.projectName, sum(a.costtodate) AS sumActivityCosts,
p.fundedbudget
From Project p inner join
Activity a
on p.projectid = a.projectid
Group By p.projectName, p.fundedbudget
having sum(a.costtodate) > p.fundedbudget;
【讨论】:
使用内部连接,我看到你为什么在查看数据后使用这种类型的连接,我收到列“Project.projectName”的错误。它是说它是无效的,因为它不包含在聚合函数或分组依据中? 找到了解决办法,我只需要在group by中添加p.projectName即可。【参考方案2】:SQL SERVER 可能适用于其他数据库
group by 下缺少字段
Group By
p.projectID
,p.fundedbudget --- to add
添加
having sum >
在 where 子句中删除 sum
更新将是
Create view vw_OverBudget AS
Select
p.projectName,
sum(a.costtodate) AS sumActivityCosts,
p.fundedbudget
From
Project AS p
full outer join Activity a ON p.projectid = a.projectid
WHERE
a.activityId is null and p.projectid is not null
Group By
p.projectID, p.fundedbudget
having SUM(a.costtodate) > p.fundedbudget
go
【讨论】:
以上是关于如何修复where子句中的聚合?的主要内容,如果未能解决你的问题,请参考以下文章
如何使用 IN 修复 Hibernate Named Native Query 的 Where 子句
如果 where 子句已经修复,如何加快 spark sql 过滤器查询?