Query to find out the time used by an user for an application

Question

Hello All,
I want to know the query to find out the whole time used by the user for an application. Please view the below data
Employee:
SNO EMP_ID EMP_NAME EMP_DATE LOGIN_TIME LOGOUT_TIME
1 10 Visu 21-Nov-2010 06:30:00 07:30:00
2 10 Visu 21-Nov-2010 06:40:00 07:20:00
3 10 Visu 21-Nov-2010 06:50:00 07:50:00
4 10 Visu 21-Nov-2010 07:30:00 08:30:00
5 10 Visu 21-Nov-2010 09:30:00 10:30:00
By checking the above data we can say that the total time Visu used the application is
8.30 - 6.30 (From 1,2,3,4 records) = 2hrs
10.30 - 9.30 (Based on 5th rec) = 1hr
So the total time Visu used the application would be 3 hrs = 180 mins.
Could you please help me in getting the result from that data using a query?

Answer 1

odie_63 wrote:
I think it may be solved with analytics too.
with t1 as (
            select 1 sno,10 emp_id,'Visu' emp_name,'21-Nov-2010' emp_date,'06:30:00' login_time,'07:30:00' logout_time from dual union all
            select 2,10,'Visu','21-Nov-2010','06:40:00','07:20:00' from dual union all
            select 3,10,'Visu','21-Nov-2010','06:50:00','07:50:00' from dual union all
            select 4,10,'Visu','21-Nov-2010','07:30:00','08:30:00' from dual union all
            select 5,10,'Visu','21-Nov-2010','09:30:00','10:30:00' from dual
     t2 as (
            select  emp_id,
                    emp_name,
                    emp_date,
                    to_date(emp_date || login_time,'DD-MON-YYYYHH24:MI:SS') login_time,
                    to_date(emp_date || logout_time,'DD-MON-YYYYHH24:MI:SS') logout_time
              from  t1
     t3 as (
            select  t2.*,
                    case
                      when login_time < max(logout_time) over(
                                                              partition by emp_id,emp_date
                                                              order by login_time
                                                              rows between unbounded preceding
                                                                       and 1 preceding
                        then 0
                      else 1
                    end start_of_group
              from  t2
     t4 as (
            select  t3.*,
                    sum(start_of_group) over(partition by emp_id,emp_date order by login_time) grp
              from  t3
     t5 as (
            select  emp_id,
                    emp_date,
                    min(login_time) login_time,
                    max(logout_time) logout_time
              from  t4
              group by emp_id,
                       emp_date,
                       grp
select  emp_id,
        numtodsinterval(sum(logout_time - login_time),'day') time_spent
  from  t5
  group by emp_id
  order by emp_id
    EMP_ID TIME_SPENT
        10 +000000000 03:00:00.000000000
SQL> SY.