Je pense que vous voulez utiliser la fonction de décalage et partitionner par dt... J'espère que cela vous aidera.
with par as (
select date '2015-08-07' enddate, 4 LookBackDays, 3600 inteval, 'Bob' Name,
'09:00' open, '14:00' close from dual),
t1 as (
select to_char(enddate-level+1, 'yyyy-mm-dd') dt, name, open, close from par
connect by level <= LookBackDays + 1 ),
t2 as (
select to_char(to_date(open, 'hh24:mi') + (level) * inteval / (24*60*60), 'hh24:mi') tm
from par
connect by to_date(open, 'hh24:mi') + level * inteval / (24*60*60)
<= to_date(close, 'hh24:mi') )
select to_date(dt, 'yyyy-mm-dd') dt,
lag(to_date(dt||' '||tm, 'yyyy-mm-dd hh24:mi'), 1,to_date(dt||' '||open, 'yyyy-mm-dd hh24:mi'))over(partition by to_date(dt||' '||open, 'yyyy-mm-dd hh24:mi') order by tm) open,
to_date(dt||' '||close, 'yyyy-mm-dd hh24:mi') close, name,
to_date(dt||' '||tm, 'yyyy-mm-dd hh24:mi') IntervalEnd
from t1 cross join t2 order by dt, tm
