问题
I have a temporal database and I want to perform the temporal aggregation via average function on time series data by using data in time sequence of following:
Hourly average of HH= (HH-1):51, HH:01, HH:11, HH:21, HH:31, HH:41.
Moreover, aforementioned values are instantaneous values.
Sample data:
ambtemp ddate_ttime
1 -1.42 2007-09-28 23:39:09
2 -1.24 2007-09-28 23:41:09
3 -1.28 2007-09-28 23:43:09
4 -1.28 2007-09-28 23:45:09
5 -1.24 2007-09-28 23:47:09
6 -1.42 2007-09-28 23:49:09
7 -1.68 2007-09-28 23:51:09
8 -1.76 2007-09-28 23:53:09
9 -1.96 2007-09-28 23:55:09
10 -2.02 2007-09-28 23:57:09
11 -1.92 2007-09-28 23:59:09
12 -1.64 2007-09-29 00:01:09
13 -1.76 2007-09-29 00:03:09
14 -1.83 2007-09-29 00:05:09
15 -1.86 2007-09-29 00:07:09
16 -1.94 2007-09-29 00:09:09
17 -1.87 2007-09-29 00:11:09
18 -1.87 2007-09-29 00:13:09
19 -1.80 2007-09-29 00:15:09
20 -1.64 2007-09-29 00:17:09
21 -1.60 2007-09-29 00:19:09
22 -1.90 2007-09-29 00:21:09
23 -2.08 2007-09-29 00:23:09
24 -1.94 2007-09-29 00:25:09
25 -2.12 2007-09-29 00:27:09
26 -1.87 2007-09-29 00:29:09
27 -2.18 2007-09-29 00:31:09
28 -1.98 2007-09-29 00:33:09
29 -1.73 2007-09-29 00:35:09
30 -1.84 2007-09-29 00:37:09
31 -2.04 2007-09-29 00:39:09
32 -1.86 2007-09-29 00:41:09
33 -1.94 2007-09-29 00:43:09
34 -1.77 2007-09-29 00:45:09
And Expected results:
Hourly mean for 0:00 = average of values of (23:51, 0:01, 0:11, 0:21, 0:31, 0:41)
Hourly mean 0:00 = (-1.68+-1.64+-1.87+-1.90+-2.18+-1.86)/6
回答1:
select dt, average_temp
from (
select
date_trunc('hour', ddate_ttime + interval '10 minutes') dt,
avg(ambtemp) over(
partition by date_trunc('hour', ddate_ttime + interval '10 minutes')
) average_temp
from sample
where extract(minute from ddate_ttime) in (51, 01, 11, 21, 31, 41)
) s
group by 1, 2
order by dt
SQL Fiddle
来源:https://stackoverflow.com/questions/14316547/average-aggregation-on-temporal-db-based-on-measurment-with-10-minute-intervals