SQL Server 2019

By using db<>fiddle, you agree to license everything you submit by Creative Commons CC0.

create table TimeLog (
  EmployeeId bigint not null,
  RoomId nchar(6) not null,
  Event nvarchar(6) not null,
  EventDate DateTime not null
)
insert TimeLog (EmployeeID, RoomId, Event, EventDate)
values (107733,	'05-27F',	'Login',	'2020-02-26 07:02:00')
     , (107733,	'05-27F',	'Logout',	'2020-02-26 08:38:00')
     , (107733,	'05-25F',	'Login',	'2020-02-26 08:39:00')
     , (107733,	'05-25F',	'Logout',	'2020-02-26 08:51:00')
     , (107733,	'05-25F',	'Login',	'2020-02-27 08:39:00')
     , (107733,	'05-25F',	'Logout',	'2020-02-27 08:51:00')
     , (107733,	'05-25X',	'Login',	'2020-02-27 08:39:00')
     , (107733,	'05-25Y',	'Logout',	'2020-02-27 08:51:00')
8 rows affected

with cte as (
  select row_number() over (partition by EmployeeId, RoomId order by EventDate) r
  , *
  from TimeLog
)
select i.EventDate LoginDate
, o.EventDate LogoutDate
, coalesce(i.EmployeeId, o.EmployeeId) EmployeeId_
  , coalesce(i.RoomId, o.RoomId)  RoomId_
from (select * from cte where Event = 'Login') i
full outer join (select * from cte where Event = 'Logout') o
on o.EmployeeId = i.EmployeeId
and o.RoomId = i.RoomId
and o.r = i.r+1
order by coalesce(i.EventDate, o.EventDate), o.EventDate, EmployeeId_, RoomId_


LoginDate
LogoutDate
EmployeeId_
RoomId_


2020-02-26 07:02:00.000
2020-02-26 08:38:00.000
107733
05-27F

2020-02-26 08:39:00.000
2020-02-26 08:51:00.000
107733
05-25F

2020-02-27 08:39:00.000
null
107733
05-25X

2020-02-27 08:39:00.000
2020-02-27 08:51:00.000
107733
05-25F

null
2020-02-27 08:51:00.000
107733
05-25Y

LoginDate	LogoutDate	EmployeeId_	RoomId_
2020-02-26 07:02:00.000	2020-02-26 08:38:00.000	107733	05-27F
2020-02-26 08:39:00.000	2020-02-26 08:51:00.000	107733	05-25F
2020-02-27 08:39:00.000	null	107733	05-25X
2020-02-27 08:39:00.000	2020-02-27 08:51:00.000	107733	05-25F
null	2020-02-27 08:51:00.000	107733	05-25Y