[英]MYSQL: select multiple if(field='x' max(time) / if(field='y' min(time)) values in a single statement
I am trying to get the below select statement to work. 我试图使下面的选择语句起作用。 Essentially I am trying to get a single row per agent & date that contains the min(time) where agent-state = 'Logged-in', the max(time) where agent-state = 'Logout', the timediff(max(time),min(time)) and the sum(duration) where agent-state = 'Ready'. 从本质上讲,我试图为每个代理&日期获取一行,其中包含min(time)其中agent-state ='Logged-in',max(time)其中agent-state ='Logout',timediff(max( time),min(time))和总和(持续时间),其中agent-state ='Ready'。
My current SQL statement is not complete to this, as I started with just getting the correct min/max time values and hit a wall. 我目前的SQL语句还不完整,因为我首先只是获得正确的最小/最大时间值,然后碰壁。 As it sits now, the Logged-in min time is returned, but not the logout max time. 现在,返回的是已登录的最短时间,但未返回注销的最长时间。
AGENT, DATE, TIME, DURATION, AGENT_STATE
Alex, 2013-01-01, 07:25:37, 00:00:00, Logged-in
Alex, 2013-01-01, 07:26:01, 00:24:48, Ready
Alex, 2013-01-01, 07:54:20, 00:21:47, Ready
Alex, 2013-01-01, 08:28:11, 00:03:00, Ready
Alex, 2013-01-01, 08:34:11, 00:06:29, Ready
Alex, 2013-01-01, 08:44:30, 00:05:14, Ready
Alex, 2013-01-01, 08:53:29, 00:06:52, Ready
Alex, 2013-01-01, 09:03:48, 00:12:13, Ready
Alex, 2013-01-01, 09:31:56, 00:36:44, Ready
Alex, 2013-01-01, 09:32:27, 14:58:51, Logout
Alex, 2013-01-01, 10:11:37, 00:00:00, Logged-in
Alex, 2013-01-01, 10:12:26, 00:54:44, Ready
Alex, 2013-01-01, 11:09:28, 00:47:48, Ready
Alex, 2013-01-01, 11:57:33, 00:16:54, Ready
Alex, 2013-01-01, 12:15:15, 00:17:32, Ready
Alex, 2013-01-01, 12:34:44, 00:13:32, Ready
Alex, 2013-01-01, 12:51:04, 00:16:44, Ready
Alex, 2013-01-01, 13:12:36, 00:31:38, Ready
Alex, 2013-01-01, 13:49:46, 00:39:14, Ready
Alex, 2013-01-01, 14:31:42, 00:28:45, Ready
Alex, 2013-01-01, 15:00:27, 14:58:51, Logout
SQL Statement: SQL语句:
select
`agent`,
`date`,
if(`agent_state` = 'Logged-in', min(`time`), null) as `min_login`,
if(`agent_state` = 'Logout', max(`time`),null) as `max_logout`
from
t_cisco_agent_state_details
group by
`agent`, `date`
order by
`agent`, `date`, `time` asc;
Thanks for any assistance or advice. 感谢您的协助或建议。
Jim 吉姆
You can't do it this way. 你不能这样子。 The reason you're getting a null for the "other" column is because that agent_state doesn't exist when you group by agent and date. 之所以对“其他”列获取空值,是因为当您按代理商和日期分组时,该agent_state不存在。 If you add the agent_state to the group-by you'll get 3 rows per agent/date. 如果将agent_state添加到分组依据,则每个代理/日期将获得3行。 This isn't quite what you want. 这不是您想要的。
You need a subquery or a couple self-joins. 您需要一个子查询或一对自联接。 Try something like this: 尝试这样的事情:
SELECT agent,
date,
MIN(dur) AS total_duration,
TIMEDIFF(MIN(maxt), MIN(mint)) AS time_difference_between_login_and_out,
MIN(maxt) AS logout_time,
MIN(mint) AS login_time
FROM (SELECT agent,
date,
IF(agent_state = 'Ready', SUM(time), NULL) AS dur,
IF(agent_state = 'Logged-in', MIN(time), NULL) AS mint,
IF(agent_state = 'Logout', MAX(time), NULL) AS maxt,
agent_state
FROM t_cisco_agent_state_details
GROUP BY agent,
date,
agent_state) AS subq
GROUP BY agent,
date
The subquery creates a single row per agent-state, agent and date. 子查询针对每个座席状态,座席和日期创建一行。 Then you can pull your times out of that with the outer query which groups by just agent and date. 然后,您可以使用外部查询(仅按代理商和日期进行分组)来节省时间。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.