朋友，
该场景与wifi控制器日志文件有关，该文件跟踪用户活动，例如当用户“加入”，“进入”，从“断开”和“离开”到wifi网络时，我已经清理并上传了结构化数据我想在用户级别为他/她的活动创建会话ID。确定一个会话可以按事件“ joins”和“ leave”之间的所有事件进行分组。为便于理解，我附加了示例数据。 / p>
一些对创建会话ID有用的指针：
请帮助我考虑给定的指针来创建会话ID。下面是示例数据，其中包含一些可能的场景，以更好地了解我要查找的内容。 “必需的数据”已分配了虚拟会话ID，这就是它的外观。张贴我的进一步澄清。我想使用SQL来实现。
usermac datetime event
xx:xx:xx:xx:0e:72 2018-04-19 09:30:58.000 joins
xx:xx:xx:xx:0e:72 2018-04-19 09:51:54.000 in
xx:xx:xx:xx:0e:72 2018-04-19 09:51:54.000 in
xx:xx:xx:xx:0e:72 2018-04-19 12:01:39.000 disconnects from
xx:xx:xx:xx:0e:72 2018-04-19 12:01:39.000 leave
xx:xx:xx:xx:0e:72 2018-04-19 12:46:05.000 joins
xx:xx:xx:xx:0e:72 2018-04-19 15:27:28.000 disconnects from
xx:xx:xx:xx:0e:72 2018-04-19 15:27:28.000 leave
usermac datetime event sessionid
xx:xx:xx:xx:0e:72 2018-04-19 09:30:58.000 joins 1
xx:xx:xx:xx:0e:72 2018-04-19 09:51:54.000 in 1
xx:xx:xx:xx:0e:72 2018-04-19 09:51:54.000 in 1
xx:xx:xx:xx:0e:72 2018-04-19 12:01:39.000 disconnects from 1
xx:xx:xx:xx:0e:72 2018-04-19 12:01:39.000 leave 1
xx:xx:xx:xx:0e:72 2018-04-19 12:46:05.000 joins 2
xx:xx:xx:xx:0e:72 2018-04-19 15:27:28.000 disconnects from 2
xx:xx:xx:xx:0e:72 2018-04-19 15:27:28.000 leave 2
usermac datetime event
xx:xx:xx:xx:xx:80 2018-04-12 21:35:48.000 joins
xx:xx:xx:xx:xx:80 2018-04-12 21:40:01.000 leave
xx:xx:xx:xx:xx:80 2018-04-12 21:40:01.000 disconnects from
xx:xx:xx:xx:xx:80 2018-04-12 21:40:08.000 joins
xx:xx:xx:xx:xx:80 2018-04-12 21:43:45.000 leave
xx:xx:xx:xx:xx:80 2018-04-12 21:43:45.000 disconnects from
xx:xx:xx:xx:xx:80 2018-04-12 21:44:00.000 joins
xx:xx:xx:xx:xx:80 2018-04-12 22:01:29.000 leave
usermac datetime event sessionid
xx:xx:xx:xx:xx:80 2018-04-12 21:35:48.000 joins 3
xx:xx:xx:xx:xx:80 2018-04-12 21:40:01.000 leave 3
xx:xx:xx:xx:xx:80 2018-04-12 21:40:01.000 disconnects from 3
xx:xx:xx:xx:xx:80 2018-04-12 21:40:08.000 joins 4
xx:xx:xx:xx:xx:80 2018-04-12 21:43:45.000 leave 4
xx:xx:xx:xx:xx:80 2018-04-12 21:43:45.000 disconnects from 4
xx:xx:xx:xx:xx:80 2018-04-12 21:44:00.000 joins 5
xx:xx:xx:xx:xx:80 2018-04-12 22:01:29.000 leave 5
usermac datetime event
xx:xx:xx:xx:xx:92 2018-04-23 09:15:18.000 joins
xx:xx:xx:xx:xx:92 2018-04-23 11:02:30.000 in
xx:xx:xx:xx:xx:92 2018-04-23 11:02:30.000 in
xx:xx:xx:xx:xx:92 2018-04-23 11:06:09.000 disconnects from
xx:xx:xx:xx:xx:92 2018-04-23 11:06:09.000 leave
xx:xx:xx:xx:xx:92 2018-04-23 11:07:03.000 joins
xx:xx:xx:xx:xx:92 2018-04-23 13:35:51.000 leave
usermac datetime event sessionid
xx:xx:xx:xx:xx:92 2018-04-23 09:15:18.000 joins 6
xx:xx:xx:xx:xx:92 2018-04-23 11:02:30.000 in 6
xx:xx:xx:xx:xx:92 2018-04-23 11:02:30.000 in 6
xx:xx:xx:xx:xx:92 2018-04-23 11:06:09.000 disconnects from 6
xx:xx:xx:xx:xx:92 2018-04-23 11:06:09.000 leave 6
xx:xx:xx:xx:xx:92 2018-04-23 11:07:03.000 joins 7
xx:xx:xx:xx:xx:92 2018-04-23 13:35:51.000 leave 7
usermac datetime event
xx:xx:xx:xx:xx:5f 2018-06-23 07:03:24.000 disconnects from
xx:xx:xx:xx:xx:5f 2018-06-23 07:03:24.000 leave
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:45.000 joins
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:48.000 leave
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:48.000 disconnects from
xx:xx:xx:xx:xx:5f 2018-06-23 07:31:59.000 joins
xx:xx:xx:xx:xx:5f 2018-06-23 07:32:02.000 leave
xx:xx:xx:xx:xx:5f 2018-06-23 07:32:02.000 disconnects from
usermac datetime event sessionid
xx:xx:xx:xx:xx:5f 2018-06-23 07:03:24.000 disconnects from 8
xx:xx:xx:xx:xx:5f 2018-06-23 07:03:24.000 leave 8
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:45.000 joins 9
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:48.000 leave 9
xx:xx:xx:xx:xx:5f 2018-06-23 07:05:48.000 disconnects from 9
xx:xx:xx:xx:xx:5f 2018-06-23 07:31:59.000 joins 10
xx:xx:xx:xx:xx:5f 2018-06-23 07:32:02.000 leave 10
xx:xx:xx:xx:xx:5f 2018-06-23 07:32:02.000 disconnects from 10
usermac datetime event
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:14.000 in
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:14.000 in
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:40.000 in
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:40.000 in
xx:xx:xx:xx:xx:b2 2018-06-12 07:23:39.000 leave
usermac datetime event sessionid
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:14.000 in 11
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:14.000 in 11
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:40.000 in 11
xx:xx:xx:xx:xx:b2 2018-06-12 07:22:40.000 in 11
xx:xx:xx:xx:xx:b2 2018-06-12 07:23:39.000 leave 11
usermac datetime event
xx:xx:xx:xx:xx:16 2018-06-23 07:55:26.000 in
xx:xx:xx:xx:xx:16 2018-06-23 07:55:26.000 in
usermac datetime event sessionid
xx:xx:xx:xx:xx:16 2018-06-23 07:55:26.000 in 12
xx:xx:xx:xx:xx:16 2018-06-23 07:55:26.000 in 12
usermac datetime event
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:30.000 rejoins
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:51.000 rejoins
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:59.000 leave
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:59.000 disconnects from
xx:xx:xx:xx:xx:m6 2018-06-24 06:57:03.000 joins
usermac datetime event sessionid
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:30.000 rejoins 13
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:51.000 rejoins 13
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:59.000 leave 13
xx:xx:xx:xx:xx:m6 2018-06-24 06:56:59.000 disconnects from 13
xx:xx:xx:xx:xx:m6 2018-06-24 06:57:03.000 joins 14
到现在为止一直在尝试：
with cte1 as
(
SELECT row_number() over(partition by usermac
ORDER BY datetime) indexno,
row_number() over(partition by event
ORDER BY usermac, datetime) sessionid, datetime, usermac, event
from t8_wfc_log_apr_jun_syslog_cast where usermac='xx:xx:xx:xx:16' order by usermac, datetime
)
SELECT case when event = 'joins' then
sessionid
else
(select max(sessionid) from cte1 innie where innie.usermac = oute.usermac and innie.indexno < oute.indexno and innie.event = 'joins') end sessionid,
datetime,
usermac,
event
FROM cte1 oute order by usermac, datetime;