使用 GROUP BY 从多个表中选择
Posted
技术标签:
【中文标题】使用 GROUP BY 从多个表中选择【英文标题】:SELECT from multiple tables with GROUP BY 【发布时间】:2012-09-16 13:49:40 【问题描述】:我在多表查询方面遇到了一点问题。 (RDBMS:访问)
这是数据库模式:(此查询中仅使用 S_Cards、Books、Authors、Students 表) S_Cards 是学生书单(在图书馆)。
查询: 选择学生中最受欢迎的作者以及该作者在图书馆订购的书籍的数量。
虽然我可以在一个查询中得到这样的订单+作者列表:
SELECT
Students.FirstName & " " & Students.LastName AS [Student],
Books.Name AS [Book], Authors.FirstName & " " & Authors.LastName AS [Author]
FROM
Students,
Books,
S_Cards,
Authors
WHERE
S_Cards.ID_Student = Students.ID
AND S_Cards.ID_Book = Books.ID
AND Books.ID_Author = Authors.ID
ORDER BY Authors.LastName
结果(抱歉,是俄语):
我不明白,为什么我不能像这样 COUNT 和 GROUP BY:
SELECT
Students.FirstName & " " & Students.LastName AS [Student],
Books.Name AS [Book],
COUNT(Authors.FirstName & " " & Authors.LastName) AS [Number of books]
FROM Students, Books, S_Cards, Authors
WHERE
S_Cards.ID_Student = Students.ID
AND S_Cards.ID_Book = Books.ID
AND Books.ID_Author = Authors.ID
GROUP BY 3
我收到一条错误消息,“Authors.FirstName & " " & Authors.LastName' 不是静态函数或组的一部分。
问题:
-
有没有办法在没有 JOIN 的情况下执行此查询,仅通过 GROUP BY、SELECT、UNION 以及如何执行?
我的第二个查询有什么问题?
【问题讨论】:
【参考方案1】:解决方案(摘自older revision in question):
SELECT TOP 1 Author, COUNT(Book) AS [Number of books] FROM
(
SELECT
Students.FirstName & " " & Students.LastName AS [Student],
Books.Name AS [Book],
Authors.FirstName & " " & Authors.LastName AS [Author]
FROM
Students,
Books,
S_Cards,
Authors
WHERE
S_Cards.ID_Student = Students.ID AND
S_Cards.ID_Book = Books.ID AND
Books.ID_Author = Authors.ID
ORDER BY Authors.LastName
)
GROUP BY Author
ORDER BY 2 DESC
【讨论】:
【参考方案2】:SQL Server 中的字符串concatenation operator 是+
而不是&
。此外,您应该按非聚合函数的事物进行分组。
SELECT Students.FirstName + ' ' + Students.LastName AS [Student]
, Books.Name AS [Book]
, COUNT(Authors.FirstName + ' ' + Authors.LastName) AS [Number of books]
FROM Students
JOIN S_Cards
ON S_Cards.ID_Student = Students.ID
JOIN BOOKS
ON S_Cards.ID_Book = Books.ID
JOIN Authors
ON Books.ID_Author = Authors.ID
GROUP BY Students.FirstName + ' ' + Students.LastName
, Books.Name
请注意,我已将您的查询更改为标准的 ANSI 连接语法,这使得错误更难产生并且更易于阅读。
仔细想想,你的计数似乎有点奇怪。书的数量不是COUNT(Books.ID)
吗?
【讨论】:
这假设您正在使用 SQL Server(看起来很像)。以后请始终使用适当的 RDBMS 标记问题。 感谢您的回复。我认为 COUNT(Books.ID) 会计算订购某本书的次数。我需要计算作者的任何书被订购了多少次,而不是特定的书。 RDBMS 是 Access。【参考方案3】:您必须对不属于聚合函数的任何内容进行 GROUP BY:
SELECT
Students.FirstName & " " & Students.LastName AS [Student],
Books.Name AS [Book],
COUNT(Authors.FirstName & " " & Authors.LastName) AS [Number of books]
FROM Students, Books, S_Cards, Authors
WHERE
S_Cards.ID_Student = Students.ID
AND S_Cards.ID_Book = Books.ID
AND Books.ID_Author = Authors.ID
GROUP BY Students.FirstName & " " & Students.LastName,
Books.Name AS [Book]
我建议您开始使用显式而不是隐式连接。在大多数情况下,MS Access 有更好的选择。
<...>
FROM Students
INNER JOIN S_Cards
ON Students.ID = S_Cards.ID_Student
或者
<...>
FROM Students
LEFT JOIN S_Cards
ON Students.ID = S_Cards.ID_Student
查询设计窗口将允许您使用正确的语法构建您的 JOIN。只需将连接字段从一个表拖放到下一个表,然后选择您需要的连接类型。
【讨论】:
以上是关于使用 GROUP BY 从多个表中选择的主要内容,如果未能解决你的问题,请参考以下文章
如何从表中选择带有 oracle sql 中的 group by 子句的嵌套 json 对象?
使用 GROUP BY,选择每个组中 ID 最低的记录 ID