Time difference by date by time range - a better way?

Question

Someone suggested this is a better place for my question:
Hi
I need to display the difference in time - every day for different time ranges,
Example 1-2 pm, 4-5 pm, 7-8 pm. And I need the time difference in 2 dates in the past week for each of these ranges everyday.
example
Date Diff Range
01/01/2007 00:01 1-2pm
01/02/2007 00:03
01/03/2007 00:10
01/04/2007 00:05
01/05/2007 00:23
01/01/2007 00:10 4-5pm
01/02/2007 00:13
01/03/2007 00:11
01/04/2007 00:15
01/05/2007 00:23
01/01/2007 01:10 7-8pm
01/02/2007 00:13
01/03/2007 00:10
01/04/2007 00:11
01/05/2007 00:21
One way to achieve this is to have multiple unions for each day and each time range.
Example:
select
from
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 13:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 14:00','mm/dd/yyyy hh24:mi:ss')
union
select
from
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 16:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 17:00','mm/dd/yyyy hh24:mi:ss')
union
select
from
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 19:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 20:00','mm/dd/yyyy hh24:mi:ss')
This will give me the required information for only one day - and that is for sysdate-5.
I will have same nyumber of unions for each day.
Is there a better way to accomplish the same?
Any help appreciated.
Thx!

Answer 1

Hi
Sorry for the late response but better late than never:::
I have gotten the answer of getting data for previous 5 days. I have changed the time between statement and is given below(*).
Here is a reply to all the questions you had asked in response to my questions.
What data you have? What parameters are you going to input?
I have already given sample data in my post.
There are no input parameters.
You are talking about the difference - between what is this difference?
Difference is the difference between 2 timestamp datatypes in 2 different tables (as you may see in the query)
The field diff - is it varchar2 like '1-2 pm' or what?
I didnt understand your question. What do you get when you subtract two timestamp datatypes - that should be the datatype - if I have understood your question. Not sure if thats what you asked.
But IMHO it's impossible to get such a result, of course, if dt_tm in the query is the same as Date in the result!
The time components in the queries are different. If you see:
1st Query:...
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 13:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 14:00','mm/dd/yyyy hh24:mi:ss')
2nd Query:....
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 16:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 17:00','mm/dd/yyyy hh24:mi:ss')
3rd Query:
where dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 19:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 20:00','mm/dd/yyyy hh24:mi:ss')
First should be between 1 and 2 pm for sysdate-5.
I need starting previous 5 days till sysdate.
Second is between 4 and 5 pm again for sysdate-5.
I need starting previous 5 days till sysdate.
Same with 3rd.
My final query is something like this:
select t1.dt_tm, count(t1.id), '1 - 3 am' as period
from table1 t1, table2 t2
where t1.id = t2.id
and dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 13:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 14:00','mm/dd/yyyy hh24:mi:ss')
group by t1.dt_tm
union
select t1.dt_tm, count(t1.id), '4 - 5 pm' as period
from table1 t1, table2 t2
where t1.id = t2.id
and dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 16:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 17:00','mm/dd/yyyy hh24:mi:ss')
group by t1.dt_tm
union
select t1.dt_tm, count(t1.id), '7 -8 pm' as period
from table1 t1, table2 t2
where t1.id = t2.id
and dt_tm between
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 19:00','mm/dd/yyyy hh24:mi:ss')and
to_date(to_char((sysdate-5),'mm/dd/yyyy')||' 20:00','mm/dd/yyyy hh24:mi:ss')
group by t1.dt_tm
I need for the last 5 days and what i can think of is have 5 different queries for past 5 days and 3 queries per day for the 3 different time periods. This would mean 15 queries. Was wondering if there is a better way to achieve the same?
Any help appreciated.
Thx!