Reputation: 2772
I've read some topics about group by sequence but I could not figure out an solution for my problem.
I have a table (the name is ViewHistory
) like this.
Tme Value
2020-07-22 09:30:00 1
2020-07-22 09:31:00 2
2020-07-22 09:32:00 3
2020-07-22 09:33:00 4
2020-07-22 09:34:00 5
2020-07-22 09:35:00 6
.
.
.
The data can grow indefinitely.
In this table, there are many records with 1 min
TimeFrame.
I want to group on range of dataTime with timeFrame 2 min
and Sum(value)
.
like this output:
TimeFrame SumData
09:30 1
09:32 5 -- sum of range 09:31_09:32
09:34 9 -- sum of range 09:33_09:34
.
.
.
How can I do this automatically, instead of using a:
WHERE Tme BETWEEN ('2020-07-22 09:31:00' AND '2020-07-22 09:32:00')
and etc.
Upvotes: 0
Views: 83
Reputation: 27201
I am sure there is a simpler way, but its not coming to me right now.
declare @Test table (tme datetime2, [value] int)
insert into @Test (tme, [value])
values
('2020-07-22 09:30:00', 1),
('2020-07-22 09:31:00', 2),
('2020-07-22 09:32:00', 3),
('2020-07-22 09:33:00', 4),
('2020-07-22 09:34:00', 5),
('2020-07-22 09:35:00', 6);
with cte as (
select convert(date, tme) [date], datepart(hour, tme) [hour], datepart(minute,dateadd(minute, 1,tme)) / 2 [minute], sum([value]) [value]
from @Test
group by convert(date, tme), datepart(hour, tme), datepart(minute,dateadd(minute, 1,tme)) / 2
)
select convert(varchar(2),[hour]) + ':' + convert(varchar(2), [minute] * 2) [time], [value]
-- , dateadd(minute, [minute] * 2, dateadd(hour, [hour], convert(datetime2, [date]))) -- Entire date if desired
from cte;
Which gives:
time | value |
---|---|
9:30 | 1 |
9:32 | 5 |
9:34 | 9 |
9:36 | 6 |
Upvotes: 1