问题描述
我有一张这样的桌子：
CREATE TABLE IF NOT EXISTS `logging` (
`id` int(6) unsigned NOT NULL,`status` varchar(150) NOT NULL,`timestamp` DATETIME NOT NULL,PRIMARY KEY (`id`)
) DEFAULT CHARSET=utf8;
INSERT INTO `logging` (`id`,`status`,`timestamp`) VALUES
('1','logout','2021-01-01 05:01:00'),('2','login','2021-01-01 06:02:00'),('3','online','2021-01-01 06:03:00'),('4','away','2021-01-01 06:04:00'),('5','2021-01-01 06:05:00'),('6','2021-01-02 04:00:00'),('7','2021-01-02 04:05:00'),('8','2021-01-02 04:07:00'),('9','2021-01-02 04:08:00'),('10','break','2021-01-02 04:10:00'),('11','2021-01-02 04:15:00'),('12','2021-01-02 04:55:00'),('13','2021-01-02 05:04:00'),('14','2021-01-02 05:05:00'),('15','2021-01-03 05:01:00'),('16','2021-01-03 05:02:00'),('17','2021-01-03 05:04:00'),('18','2021-01-03 05:05:00'),('19','2021-01-04 03:05:00'),('20','2021-01-04 05:07:00'),('21','2021-01-04 06:00:00'),('22','2021-01-05 10:00:00'),('23','2021-01-05 11:00:00'),('24','2021-01-05 11:01:00'),('25','2021-01-06 06:01:00'),('26','2021-01-07 06:01:00'),'2021-01-07 07:01:00');
| id | 状态 | 时间戳 |
|---|---|---|
| 1 | 注销 | 2021-01-01 05:01:00 |
| 2 | 登录 | 2021-01-01 06:02:00 |
| 3 | 在线 | 2021-01-01 06:03:00 |
| 4 | 离开 | 2021-01-01 06:04:00 |
| 5 | 在线 | 2021-01-01 06:05:00 |
| 6 | 注销 | 2021-01-02 04:00:00 |
| 7 | 登录 | 2021-01-02 04:05:00 |
| 8 | 在线 | 2021-01-02 04:07:00 |
| 9 | 注销 | 2021-01-02 04:55:00 |
......
上面插入查询中提供的数据。
| 日期 | A（在线） | B（注销） |
|---|---|---|
| 2021-01-01 | 2021-01-02 04:07:00 | 2021-01-02 04:55:00 |
| 2021-01-02 | 2021-01-02 05:05:00 | 2021-01-03 04:59:59 |
| 2021-01-03 | 2021-01-03 05:05:00 | 2021-01-04 03:05:00 |
| 2021-01-04 | 2021-01-04 06:00:00 | 2021-01-04 04:59:59 |
| 2021-01-05 | 2021-01-04 05:00:00 | 2021-01-05 10:00:00 |
| 2021-01-06 | 2021-01-04 11:00:00 | 2021-01-06 04:59:59 |
规则是，1 个日志日是从 5:00:00 -（第二天）04:59:59。 'A' 是从最后一次在线开始的时间戳（登录后，如果有的话），'B' 是从最后一次注销开始的时间戳（如果没有注销，B 设置为 04:59:59。 另一个规则是，当最后一天在 'online' 之后没有 'logout' 时，将计入下一个日志日（如果最后一个日志日的最后一条记录为 '，则下一个日志日设置为 05.00.00在线')
目前我正在使用此查询来应用日志日规则：
SELECT date(t1.timestamp) dt,(t2.timestamp) A,(t3.timestamp) B,t1.status,t2.status,t3.status
FROM logging t1
JOIN logging t2 ON t1.timestamp < t2.timestamp
JOIN logging t3 ON t2.timestamp < t3.timestamp
WHERE
t1.status = 'login'
AND t2.status = 'online'
AND t3.status = 'logout'
AND NOT EXISTS ( SELECT NULL
FROM loggingt4
WHERE t1.timestamp < t4.timestamp
AND t4.timestamp < t2.timestamp
AND t4.status IN ('login','logout') )
AND NOT EXISTS ( SELECT NULL
FROM logging t5
WHERE t2.timestamp < t5.timestamp
AND t5.timestamp < t3.timestamp
AND t5.status IN ('login','logout'))
AND DATE(t1.timestamp - INTERVAL 6 HOUR) = DATE(t3.timestamp - INTERVAL '05:59:59' HOUR_SECOND);
解决方法
检查这个：
WITH cte AS (
SELECT DATE(t1.`timestamp` - INTERVAL 5 HOUR) `date`,MAX(t1.`timestamp`) login,MAX(t2.`timestamp`) online,MAX(t3.`timestamp`) logout
FROM logging t1
JOIN logging t2 ON t1.`timestamp` < t2.`timestamp`
JOIN logging t3 ON t2.`timestamp` < t3.`timestamp`
WHERE t1.status = 'login'
AND t2.status = 'online'
AND t3.status = 'logout'
AND NOT EXISTS ( SELECT NULL
FROM logging t4
WHERE t1.`timestamp` < t4.`timestamp`
AND t4.`timestamp` < t2.`timestamp`
AND t4.status IN ('login','online','logout') )
AND NOT EXISTS ( SELECT NULL
FROM logging t5
WHERE t2.`timestamp` < t5.`timestamp`
AND t5.`timestamp` < t3.`timestamp`
AND t5.status IN ('login','logout') )
GROUP BY `date`
)
SELECT `date`,online,CASE WHEN DATE(online - INTERVAL 5 HOUR) = DATE(logout - INTERVAL '04:59:59' HOUR_SECOND)
THEN logout
ELSE DATE(online + INTERVAL 19 HOUR) + INTERVAL '04:59:59' HOUR_SECOND
END logout
FROM cte
CTE 仅用于可见性 - 您可以将所有内容合并为一个查询。