在此查询中获取数据计数的更好方法？ [复制]

Posted 2023-03-24

【中文标题】在此查询中获取数据计数的更好方法？ [复制]

【英文标题】：Better way to get counts of data in this query? [duplicate]

【发布时间】：2018-07-06 19:12:41

【问题描述】：

我有一个图书馆 (books) 和另一个桌子，每次有人借书或带回一本书时，它都会被记录下来 (book_loans)。

基本上我想查看在某个时间签出的所有书籍，有多少：

属于“奇幻”类型 超过 500 页 是精装书

我想一次完成所有这些，我现在使用 4 个查询，有没有更好的方法？

PS：book_loans.bookID 是books.id 的外键：

SELECT * FROM
(

    SELECT COUNT(*) FROM `books` AS bks
    INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
    WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1 AND bks.genre = 'fantasy'

    UNION

    SELECT COUNT(*) FROM `books` AS bks
    INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
    WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1 AND bks.pages > 500


    UNION

    SELECT COUNT(*) FROM `books` AS bks
    INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
    WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1 AND bks.hardcover = 1

) x;

为此总共执行 4 个查询似乎效率很低。

【参考方案1】：

SELECT SUM(bks.genre = 'fantasy') as fantasy,
       SUM(bks.pages > 500) as pages,
       SUM(bks.hardcover = 1) as hardcover
FROM `books` AS bks
INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1

【讨论】：

非常感谢！

【参考方案2】：

如果您只想要全部 3 个的总数，或者，如果您需要 3 个不同的总数，每个总数一个，或者您需要不同的东西，这应该可以满足您的需要。

SELECT COUNT(*) FROM `books` AS bks
INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1 
AND (
    bks.genre = 'fantasy'
    OR bks.pages > 500
    OR bks.hardcover = 1
)

【参考方案3】：

SELECT CASE WHEN bks.genre = 'fantasy' THEN 'genre match' 
            WHEN bks.pages > 500 THEN 'page count met' 
            WHEN bks.hardcover = 1 THEN 'is hardcover'
       END AS condition
      , COUNT(*) 
FROM `books` AS bks
INNER JOIN `book_loans` AS bs ON bks.id = bs.bookID
WHERE bs.time = '2017-07-01 02:16:00' AND bs.signedout = 1 
  AND (bks.genre = 'fantasy' OR bks.pages > 500 OR bks.hardcover = 1)
GROUP BY condition;

这是 juergen d's 的替代方法，具有相似的结果；但我会先和他一起去。与他不同的是，在这本书中，匹配多个条件的书只会包含在第一个条件的计数中。

但是，如果 book_loans 表非常大，您仍可以考虑在 WHERE 子句中包含 OR 条件。