Question

我有下表：

CREATE TABLE IF NOT EXISTS `access_log` (
    `id` INT(11) NOT NULL AUTO_INCREMENT,
    `user_id` INT(11) NOT NULL DEFAULT 0,
    `room_id` INT(11) NOT NULL DEFAULT 0,
    `created` TIMESTAMP NOT NULL,
    PRIMARY KEY (`id`)
) ENGINE=MyISAM  DEFAULT CHARSET=utf8;

每次用户进入房间room_id都会添加一条新记录。我想为每个房间选择每个用户的第一条记录和最后一条记录。

目前，我有以下查询似乎未提供正确的记录：

对于每个用户的每个房间的第一记录：

SELECT al.* FROM `access_log` AS `al`
LEFT JOIN `rooms` AS `r` ON al.room_id = r.id
INNER JOIN (
    SELECT user_id, room_id, min(created) AS min_date
    FROM `access_log`
    WHERE `user_id` != 0
    GROUP BY user_id, room_id) AS al2
ON al.user_id = al2.user_id AND al.room_id = al2.room_id AND al.created = al2.min_date
WHERE `al`.`created` >= '2019-06-09 00:00:00' AND `al`.`created` <= '2019-06-12 23:59:59'

对于每个用户的每个房间的最后记录：

SELECT al.* FROM `access_log` AS `al`
LEFT JOIN `rooms` AS `r` ON al.room_id = r.id
INNER JOIN (
    SELECT user_id, room_id, max(created) AS max_date
    FROM `access_log`
    WHERE `user_id` != 0
    GROUP BY user_id, room_id) AS al2
ON al.user_id = al2.user_id AND al.room_id = al2.room_id AND al.created = al2.max_date
WHERE `al`.`created` >= '2019-06-09 00:00:00' AND `al`.`created` <= '2019-06-12 23:59:59'

这是一个SQLFiddle演示，其中包括示例数据http://www.sqlfiddle.com/#!9/fc5f8b/2。您可以看到查询显示意外的结果。他们没有列出不同的日期，尽管他们列出了不同的房间。另外，第一个查询和最后一个查询的行数也不同。

相同的DDL：

CREATE TABLE IF NOT EXISTS `access_log` (
    `id` INT(11) NOT NULL AUTO_INCREMENT,
    `user_id` INT(11) NOT NULL DEFAULT 0,
    `room_id` INT(11) NOT NULL DEFAULT 0,
    `created` TIMESTAMP NOT NULL,
    PRIMARY KEY (`id`)
) ENGINE=MyISAM  DEFAULT CHARSET=utf8;

INSERT INTO `access_log` (`id`, `user_id`, `room_id`, `created`) VALUES
(1, 90000017, 6, '2019-06-10 01:15:00'),
(2, 90000017, 6, '2019-06-10 01:25:00'),
(3, 90000018, 6, '2019-06-10 02:15:00'),
(4, 90000018, 6, '2019-06-10 02:25:00'),
(5, 90000019, 6, '2019-06-10 03:15:00'),
(6, 90000019, 6, '2019-06-10 03:25:00'),

(7, 90000017, 5, '2019-06-10 11:15:00'),
(8, 90000017, 5, '2019-06-10 11:25:00'),
(9, 90000018, 5, '2019-06-10 12:15:00'),
(10, 90000018, 5, '2019-06-10 12:25:00'),
(11, 90000019, 5, '2019-06-10 13:15:00'),
(12, 90000019, 5, '2019-06-10 13:25:00'),

(13, 90000017, 6, '2019-06-11 04:10:00'),
(14, 90000017, 6, '2019-06-11 04:20:00'),
(15, 90000018, 6, '2019-06-11 05:10:00'),
(16, 90000018, 6, '2019-06-11 05:20:00'),
(17, 90000019, 6, '2019-06-11 06:10:00'),
(18, 90000019, 6, '2019-06-11 06:20:00'),

(19, 90000017, 5, '2019-06-11 14:10:00'),
(20, 90000017, 5, '2019-06-11 14:20:00'),
(21, 90000018, 5, '2019-06-11 15:10:00'),
(22, 90000018, 5, '2019-06-11 15:20:00'),
(23, 90000019, 5, '2019-06-11 16:20:00'),
(24, 90000019, 5, '2019-06-11 16:20:00');

预期结果应该类似于：

First per user per room per day
+------+-----------+---------+---------------------+
|  id  |  user_id  | room_id |       created       |
+------+-----------+---------+---------------------+
|  1   | 90000017  |    6    | 2019-06-10 01:15:00 |
|  3   | 90000018  |    6    | 2019-06-10 02:15:00 |
|  5   | 90000019  |    6    | 2019-06-10 03:15:00 |
|  7   | 90000017  |    5    | 2019-06-10 11:15:00 |
|  9   | 90000018  |    5    | 2019-06-10 12:15:00 |
|  11  | 90000019  |    5    | 2019-06-10 13:15:00 |
|  13  | 90000017  |    6    | 2019-06-11 04:10:00 |
|  15  | 90000018  |    6    | 2019-06-11 05:10:00 |
|  17  | 90000019  |    6    | 2019-06-11 06:10:00 |
|  19  | 90000017  |    5    | 2019-06-11 14:10:00 |
|  21  | 90000018  |    5    | 2019-06-11 15:10:00 |
|  23  | 90000019  |    5    | 2019-06-11 16:20:00 |
+------+-----------+---------+---------------------+

Last per user per room per day
+------+-----------+---------+---------------------+
|  id  |  user_id  | room_id |       created       |
+------+-----------+---------+---------------------+
|  2   | 90000017  |    6    | 2019-06-10 01:25:00 |
|  4   | 90000018  |    6    | 2019-06-10 02:25:00 |
|  6   | 90000019  |    6    | 2019-06-10 03:25:00 |
|  8   | 90000017  |    5    | 2019-06-10 11:25:00 |
|  10  | 90000018  |    5    | 2019-06-10 12:25:00 |
|  12  | 90000019  |    5    | 2019-06-10 13:25:00 |
|  14  | 90000017  |    6    | 2019-06-11 04:20:00 |
|  16  | 90000018  |    6    | 2019-06-11 05:20:00 |
|  18  | 90000019  |    6    | 2019-06-11 06:20:00 |
|  20  | 90000017  |    5    | 2019-06-11 14:20:00 |
|  22  | 90000018  |    5    | 2019-06-11 15:20:00 |
|  24  | 90000019  |    5    | 2019-06-11 16:20:00 |
+------+-----------+---------+---------------------+

Answer 1

我建议通过以下单个查询对预期结果进行交叉检查：

SELECT 
   GROUP_CONCAT(id ORDER BY created,id SEPARATOR ' ') all_id, 
   -- this return all id present in the group
   SUBSTRING_INDEX(GROUP_CONCAT(id ORDER BY created,id SEPARATOR ' '),' ',1) min_id_in,
   -- this part is taking the first value from the GROUP_CONCAT operation above
   SUBSTRING_INDEX(GROUP_CONCAT(id ORDER BY created,id SEPARATOR ' '),' ',-1) max_id_in,
   -- this part is taking the last value from the first GROUP_CONCAT operation
   user_id,room_id,
   MIN(created), 
   MAX(created) -- Min/max value are both shown in same query
FROM access_log 
GROUP BY user_id,room_id,
date(created); -- the missing condition where OP's asks results to return by each date.

我在date(created)条件下添加了GROUP BY ..。

在小提琴的原始查询中：

SELECT al.* FROM `access_log` AS `al`
INNER JOIN (
    SELECT user_id, room_id, min(created) AS min_date
    FROM `access_log`
    WHERE `user_id` != 0
    GROUP BY user_id, room_id, 
    date(created) -- I've added the condition here 
    ) AS al2
ON al.user_id = al2.user_id AND al.room_id = al2.room_id AND al.created = al2.min_date
WHERE `al`.`created` >= '2019-06-09 00:00:00' AND `al`.`created` <= '2019-06-12 23:59:59'
ORDER BY al.user_id ASC;


SELECT al.* FROM `access_log` AS `al`
INNER JOIN (
    SELECT user_id, room_id, max(created) AS max_date
    FROM `access_log`
    WHERE `user_id` != 0
    GROUP BY user_id, room_id, 
    date(created) -- and here
    ) AS al2
ON al.user_id = al2.user_id AND al.room_id = al2.room_id AND al.created = al2.max_date
WHERE `al`.`created` >= '2019-06-09 00:00:00' AND `al`.`created` <= '2019-06-12 23:59:59'
ORDER BY al.user_id ASC;

Answer 2

子查询应从与主查询相同的表中进行选择，因此应从del而非rm /outbound/test/FileName_A_* rm /outbound/test/FileName_B_*中进行选择。

access_log

如何获取MySQL中每个用户的第一条记录和最后一条记录（用于出勤）

2 个答案: