I posted question but still haven't got problem solved, please help

Question

I have a table which has two columes
time- timestamp
value- integer
I have problem writing query to get rolling up cumulative of value for time buckets
Wanted results like following:
like cumulative value for rolling 5 minutes time bucket for past 1 hour
If current time is 8:51:00
value cumulative value
8:51:00 3 256
8:46:00 4 253
8:01:00 3 14
7:56:00 6 11
7:51:00 5 5
user want the time bucket can set to 5min, 1 hours, 1 day
and the data can be refreshed every 1 min. Is there some way we can get live rolling time bucket like
if currenct time is 3/18/2006 11:07:23 am, the interval(Time bucket) send to 1 hour. Results like following
count(*) interval interval3
187 88 3/18/2006 10:07:00 AM
280 87 3/18/2006 09:07:00 AM
34 86 3/18/2006 08:07:00 AM
1 mins later. time is 3/18/2006 11:07:23 am. refresh data, the result like
count(*) interval interval3
188 88 3/18/2006 10:08:00 AM
282 87 3/18/2006 09:08:00 AM
36 86 3/18/2006 08:08:00 AM
Is it possible?
Thanks a lot
previous question link
Re: Help about rolling cumulative sum in time bucket

Answer 1

The referenced thread had a much better subject line.
Question can the number of required buckets that are brought back be limited to say 12? If you can limit the result set to a fixed number of buckets then you should be able to just use a group by on a case statement with a little date/time math in it. That is you would be able to calculate the time boundries from the present and then compare the row data timestamps to the boundries.
With a variable number of buckets you would need to find the difference in the desired units (5 min, 1 hour, days ? 24 hours) between the current_timestamp and the oldest timestamp in the data. You would then need to calculate the boundries and compare the data. Doing all this would be more resource intensive than the group by the case test.
HTH -- Mark D Powell --