Calculate how long a process took, taking into account opening hours - sql

Question

I have two tables. An opening hours table that says for each seller and store, which are the opening and closing times for each day of the week. The second table is the operation one which has all information about the processes.
What I need is to calculate how many seconds each process took considering only the hours when the store was opened.
I tried to solve that with case when. I solved the problem when the process take less than 2 days. But I don't know how to handle it when it takes more days. The other problem I had with this code is that case when takes a lot of time to process. Can anybody help me with these issues?
Opening hours table:
sellerid
sellerstoreid
day
dayweek
opening
closing
next_day
opening_next_day
days_to_next
123
abc
1
monday
09:00:00
17:00:00
2
09:00:00
1
123
abc
2
tuesday
09:00:00
17:00:00
4
09:00:00
2
123
abc
4
thursday
09:00:00
17:00:00
5
09:30:00
1
123
abc
5
friday
09:30:00
17:00:00
1
09:00:00
3
Where:
sellerid + sellerstoreid + day works as a primary key;
dayweek translates day from number to name;
opening and closing are the opening and closing time for that day;
opening_next_day shows the opening time o the next available date for that store and seller;
days_to_next informes in how many days will the store reopen
Process table:
delivery_id
sellerid
sellerstoreid
process
end_time
a1
123
abc
p1
05/12/2022 16:00:00.000
a1
123
abc
p2
06/12/2022 16:00:00.000
a1
123
abc
p3
06/12/2022 16:00:00.000
a1
123
abc
p4
08/12/2022 16:00:00.000
a1
123
abc
p5
13/12/2022 16:00:00.000
Where:
The end_time of the previous process will be the the start time of the process.
with
table_1 as (
select
delivery_id
, sellerid
, sellerstoreid
, process
, lag(end_time, 1) over (partition by delivery_id order by end_time) as start_time
, extract(dow from lag(end_time, 1) over (partition by delivery_id order by end_time)) as dow_start_time
, end_time
, extract(dow from end_time) as dow_end_time
from process_table
),
table_2 as (
select
table_1.*
, oh_start.opening as start_opening
, oh_start.closing as start_closing
, oh_end .opening as end_opening
, oh_end .closing as end_closing
from table_1 tb1
left join opening_hours oh_start
on oh_start.sellerid = tb1.sellerid
and oh_start.sellerstoreid = tb1.sellerstoreid
and oh_start.day = dow_start_time
left join opening_hours oh_end
on oh_end .sellerid = tb1.sellerid
and oh_end.sellerstoreid = tb1.sellerstoreid
and oh_end.day = dow_end_time
)
select
*
, case
when dow_start_time = dow_end_time then
extract(epoch from
(case
when end_time::time > start_opening then
(case
when end_time::time > start_closing then start_closing
else end_time::time
end)
else start_opening
end
-
case
when start_time::time > start_opening then
(case
when start_time::time < start_closing then start_time::time
else start_closing
end
)
else start_opening
end))
when dow_start_time <> dow_end_time then
extract(epoch from
(start_closing
-
case
when start_time::time > start_opening then
(case
when start_time::time < start_closing then start_time::time
else start_closing
end)
else start_opening
end)
+
(case
when end_time::time > end_opening then
(case
when end_time::time > end_closing then end_closing
else end_time::time
end)
else end_opening
end
-
end_opening)
end status_duration
from table_2