如何选择每组的前N行?

时间:2022-03-02 09:09:08

I have two SQLite tables like this:

我有两个这样的SQLite表:

 AuthorId | AuthorName
----------------------
 1        | Alice
 2        | Bob
 3        | Carol
 ...      | ....


 BookId | AuthorId | Title
----------------------------------
 1      | 1        | aaa1
 2      | 1        | aaa2
 3      | 1        | aaa3
 4      | 2        | ddd1
 5      | 2        | ddd2
 ...    | ...      | ...
 19     | 3        | fff1
 20     | 3        | fff2
 21     | 3        | fff3
 22     | 3        | fff4

I want to make a SELECT query that will return the first N (e.g. two) rows for each AuthorId, ordering by Title ("Select the first two books of each author").

我想创建一个SELECT查询,它将返回每个AuthorId的前N行(例如两行),按Title排序(“选择每个作者的前两本书”)。

Sample output:

样本输出:

 BookId |  AuthorId | AuthorName | Title
------------------------------------------
 1      |  1        |   Alice    | aaa1
 2      |  1        |   Alice    | aaa1
 4      |  2        |   Bob      | ddd1
 5      |  2        |   Bob      | ddd2
 19     |  3        |   Carol    | fff1
 20     |  3        |   Carol    | fff2

How can I build this query?

如何构建此查询?

(Yes, I found a similar topic, and I know how to return only one row (first or top). The problem is with the two).

(是的,我发现了一个类似的主题,我知道如何只返回一行(第一行或顶部)。问题在于两者)。

3 个解决方案

#1


14  

You can do the counting using a correlated subquery:

您可以使用相关子查询进行计数:

SELECT b.BookId, a.AuthorId, a.AuthorName, b.Title
FROM Author a join
     Book b
     on a.AuthorId = b.AuthorId
where (select count(*)
       from book b2
       where b2.bookId <= b.BookId and b2.AuthorId = b.AuthorId
      ) <= 2;

For a small database this should be fine. If you create a composite index on Book(AuthorId, BookId) then that will help the query.

对于一个小型数据库,这应该没问题。如果您在Book(AuthorId,BookId)上创建复合索引,那么这将有助于查询。

#2


13  

There is alternative variant:

还有另一种变体:

SELECT * FROM (
    SELECT * FROM BOOK, AUTHOR
    WHERE BOOK.AUTHORID = AUTHOR.AUTHORID
) T1
WHERE T1.BOOKID IN (
    SELECT T2.BOOKID FROM BOOK T2
    WHERE T2.AUTHORID = T1.AUTHORID
    ORDER BY T2.BOOKTITLE
    LIMIT 2
)
ORDER BY T1.BOOKTITLE

#3


0  

Here you go. Might be too late but I just saw the post. You can change the <=2 to match the n you need.

干得好。可能为时已晚,但我刚看到帖子。您可以更改<= 2以匹配您需要的n。

SELECT 
 a.authorid,
 a.authorname,
 b.bookid,
 b.booktitle
FROM author a
JOIN book b ON b.authorid = b.authorid
QUALIFY ROW_NUMBER() OVER (PARTITION BY a.authorid
ORDER BY   b.booktitle ASC) <=2

#1


14  

You can do the counting using a correlated subquery:

您可以使用相关子查询进行计数:

SELECT b.BookId, a.AuthorId, a.AuthorName, b.Title
FROM Author a join
     Book b
     on a.AuthorId = b.AuthorId
where (select count(*)
       from book b2
       where b2.bookId <= b.BookId and b2.AuthorId = b.AuthorId
      ) <= 2;

For a small database this should be fine. If you create a composite index on Book(AuthorId, BookId) then that will help the query.

对于一个小型数据库,这应该没问题。如果您在Book(AuthorId,BookId)上创建复合索引,那么这将有助于查询。

#2


13  

There is alternative variant:

还有另一种变体:

SELECT * FROM (
    SELECT * FROM BOOK, AUTHOR
    WHERE BOOK.AUTHORID = AUTHOR.AUTHORID
) T1
WHERE T1.BOOKID IN (
    SELECT T2.BOOKID FROM BOOK T2
    WHERE T2.AUTHORID = T1.AUTHORID
    ORDER BY T2.BOOKTITLE
    LIMIT 2
)
ORDER BY T1.BOOKTITLE

#3


0  

Here you go. Might be too late but I just saw the post. You can change the <=2 to match the n you need.

干得好。可能为时已晚,但我刚看到帖子。您可以更改<= 2以匹配您需要的n。

SELECT 
 a.authorid,
 a.authorname,
 b.bookid,
 b.booktitle
FROM author a
JOIN book b ON b.authorid = b.authorid
QUALIFY ROW_NUMBER() OVER (PARTITION BY a.authorid
ORDER BY   b.booktitle ASC) <=2