Combine multiple rows of data into one (Start and End time)

Maybe a group by query like below on row_number column

see fiddle link

select objectID, 
       min(msgTime) as OnTime, 
       case 
         when min(MsgTime) <>max(MsgTime) 
         then max(MsgTime) else NULL 
end as OffTime 
from 
(
   select *, 
  row_number() over (partition by ObjectID order by MsgTime asc)+1 as r 
  from T
 )T
group by objectID, r/2
order by Objectid, r/2

This query would return all 'going to on state' rows, and for each one it finds the nearest 'going to off state' row, if exists (LEFT JOIN)

select 
   ontm.ObjectID, ontm.MsgTime as OnTime, offtm.MsgTime as OffTime
from yourtable ontm
     left join
     yourtable offtm
     on ontm.ObjectId=offtm.ObjectId
        and offtm.onoffid = 0
        and ontm.MsgTime <= offtm.MsgTime
        and not exists (select 1
                        from yourtable mdle
                        where mdle.ObjectId=offtm.ObjectId
                          and mdle.MsgTime < offtm.MsgTime
                          and ontm.MsgTime < mdle.MsgTime  
                       )  
where ontm.onoffid = 1

Explanation: We first select all 'going to on' rows; these are the ones we want a result row. We then find all 'matching', i.e. future, 'going to off' state records for the same objectid (we use LEFT JOIN to make sure that if the objectId was left ON we still show it). This would match all future 'going to off' rows for the object, so we need something to make sure that only the earliest one matches our ON row; we do this by making sure that for any candidate OFF row, there is no other future, earlier OFF row: NOT EXISTS.