如何在MySQL中优化多个左连接？

Question

你能帮我让这个查询正常工作吗？

我有一个日志查询，它按天(表：log，包括没有行的天数)计算每个活动用户(表：user，状态：1 (表示活动))的已记录项目(表：calendar )。

以下查询需要10分钟才能运行！如何在几秒而不是几分钟内运行此程序？

SELECT
    c.day, COUNT(u.id) AS count
FROM calendar c
LEFT JOIN log l
    ON c.day = DATE_FORMAT(l.db_timestamp , '%Y-%m-%d')
LEFT JOIN user u
    ON l.user_id = u.id
    AND u.user_status_type_id = 1
WHERE
    c.day > '2012-12-01'
    AND c.day < '2013-01-01'
GROUP BY
    c.day

表结构：

calendar (~3,000 rows)
day
===============================
2012-01-01
2012-01-02
2012-01-03
...
2020-01-01


log (~30,000 rows)
id  user_id  db_timestamp
================================
1   1        2012-01-01 01:01:01
1   2        2012-01-01 01:01:01
1   1        2012-01-01 01:01:01

user (~3,000,000 rows)
id  user_status_type_id
================================
1   1
1   0

结果应如下所示：

Sample Expected Results
day         count
=================
2012-12-01  1
2012-12-02  0
2012-12-03  4
...
2012-12-31  0

不幸的是，它需要永远运行。我接下来该怎么办？

Answer 1

对于您选择的列，您不需要任何连接。使用以下sql

SELECT DATE_FORMAT(l.db_timestamp , '%Y-%m-%d') AS days, COUNT(l.id) AS COUNT 
FROM LOG l 
WHERE 
DATE_FORMAT(l.db_timestamp , '%Y-%m-%d') > '2012-12-01' 
AND DATE_FORMAT(l.db_timestamp , '%Y-%m-%d') < '2013-01-01'
GROUP BY days

对于用户智能计数

使用

GROUP BY days, l.user_id

Answer 2

尝试此操作：：在加入时使用DATE()

 SELECT
        c.day, COUNT(u.id) AS count
    FROM calendar c
    LEFT JOIN log l
        ON c.day = DATE(l.db_timestamp)
    LEFT JOIN user u
        ON l.user_id = u.id
        AND u.user_status_type_id = 1
    WHERE
        c.day between '2013-01-01'
 AND '2012-12-01'
        GROUP BY
        c.day