Question

SELECT
b.cID,
b.ID,
Count(r.userRead) AS readCount,
COUNT(DISTINCT r.userID) AS UserCount,
Count(c.userDownload) AS downloadCount,
COUNT(DISTINCT c.userID) AS userDownloadCount
FROM
 book AS b
  INNER JOIN book_event AS r ON r.bookID=s.ID AND r.bookRead = 1
  INNER JOIN book_event as c ON c.bookID=s.ID AND c.bookDownload = 1
WHERE
 b.cID = 1011
GROUP BY
 b.ID
ORDER BY
 b.ID DESC

此SQL查询输出（计数问题）

+-----------+-----+-----------+-----------------+--------------+-------------------+
| cID       | ID  | readCount | UserCount       | downloadCount| userDownloadCount |
+-----------+-----+-----------+-----------------+--------------+-------------------+
|      1011 | 278 |      3168 |              67 |         3168 |                19 |
|      1011 | 272 |      9918 |             122 |         9918 |                41 |
|      1011 | 241 |     31694 |              99 |        31694 |                38 |
+-----------+-----+-----------+-----------------+--------------+-------------------+
3 rows in set

实际价值

+-----------+-----+-----------+-----------------+--------------+-------------------+
| cID       | ID  | readCount | UserCount       | downloadCount| userDownloadCount |
+-----------+-----+-----------+-----------------+--------------+-------------------+
|      1011 | 278 |       133 |              67 |           24 |                19 |
|      1011 | 272 |       174 |             122 |           57 |                41 |
|      1011 | 241 |       299 |              99 |          106 |                38 |
+-----------+-----+-----------+-----------------+--------------+-------------------+

book_event（table）

+-----+--------+----------+--------------+
| ID  | userID | userRead | userDownload |
+-----+--------+----------+--------------+
| 278 |   5169 | 1        | 0            |
| 278 |   5169 | 0        | 1            |
| ... |   .... | .        | .            |
| 278 |   5628 | 1        | 0            |
| 278 |   5162 | 1        | 0            |
+-----+--------+----------+--------------+

我需要将计数分为两列。 readCount和downloadCount columuns不正确，但UserCount，userDownloadCount columuns值是正确的。

我该如何解决这个问题？

Answer 1

这是因为您在同一本书上有多个读取和下载事件，因此您的查询会生成事件的交叉产品。

解决此问题的一个好方法是分别汇总信息。但是，您的查询提供了更简单的解决方案只需加入book_event表一次，然后计算不同的事件。

SELECT b.cID, b.ID,
       sum(be.bookRead) as readCount,
       count(distinct case when be.bookRead = 1 then be.userId end) as UserCount,
       sum(be.userDownload) as downloadCount,
       count(distinct case when be.userDownload = 1 then be.userId end) as userDownloadCount
FROM book AS b INNER JOIN
     book_event be
     on be.bookID = s.ID
WHERE b.cID = 1011
GROUP BY b.ID, b.cid
ORDER BY b.ID DESC

我将b.cid添加到group by子句中。在GROUP BY中的SELECT子句中包含所有非聚合值是一种很好的形式。其他数据库强制执行此操作，该规则是标准SQL。

Answer 2

我认为您的查询在技术上是错误的（按语法分组），但它与您的问题无关并适用于MySQL。

计数问题的最常见原因是没有正确计算空值 - 但我不认为这也是你的问题。

试试这个

SELECT 
b.cID, 
b.ID, 
sum(distinct coalesce(r.userRead,0)) AS readCount, 
count(DISTINCT r.userID) AS UserCount, 
sum(distinct coalesce(c.userDownload,0)) AS downloadCount, 
count(DISTINCT c.userID) AS userDownloadCount 
FROM 
 book AS b 
  left JOIN book_event AS r ON r.bookID=s.ID AND r.bookRead = 1 
  left JOIN book_event as c ON c.bookID=s.ID AND c.bookDownload = 1 
WHERE 
 b.cID = 1011 
GROUP BY 
 b.cID, b.ID 
ORDER BY 
 b.ID DESC

MYSQL计数不正确

2 个答案: