Average of the values grouped over a period of time - sql-server

Question

I want to find the average of the grouped values
with the rule (if time(n+1) - time(n) < 40 minutes) and by feature
<style type="text/css">
.tg {border-collapse:collapse;border-spacing:0;}
.tg td{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
overflow:hidden;padding:10px 5px;word-break:normal;}
.tg th{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
font-weight:normal;overflow:hidden;padding:10px 5px;word-break:normal;}
.tg .tg-0pky{border-color:inherit;text-align:left;vertical-align:top}
</style>
<table class="tg">
<thead>
<tr>
<th class="tg-0pky">feature</th>
<th class="tg-0pky">value</th>
<th class="tg-0pky">time</th>
</tr>
</thead>
<tbody>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,76</td>
<td class="tg-0pky">20200508 11:13:06</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,69</td>
<td class="tg-0pky">20200508 11:13:13</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,94</td>
<td class="tg-0pky">20200508 11:13:20</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,73</td>
<td class="tg-0pky">20200508 11:13:27</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,79</td>
<td class="tg-0pky">20200508 12:16:55</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,76</td>
<td class="tg-0pky">20200508 12:17:03</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">60,1</td>
<td class="tg-0pky">20200508 12:17:10</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">59,95</td>
<td class="tg-0pky">20200508 12:17:18</td>
</tr>
<tr>
<td class="tg-0pky">Diameter</td>
<td class="tg-0pky">60,02</td>
<td class="tg-0pky">20200508 12:17:36</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,1</td>
<td class="tg-0pky">20200508 05:23:30</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,08</td>
<td class="tg-0pky">20200508 05:23:35</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,04</td>
<td class="tg-0pky">20200508 05:23:40</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,06</td>
<td class="tg-0pky">20200508 05:23:46</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,16</td>
<td class="tg-0pky">20200508 05:23:52</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,69</td>
<td class="tg-0pky">20200508 06:03:05</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,65</td>
<td class="tg-0pky">20200508 06:03:13</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,64</td>
<td class="tg-0pky">20200508 06:03:18</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,39</td>
<td class="tg-0pky">20200508 06:03:24</td>
</tr>
<tr>
<td class="tg-0pky">Weight</td>
<td class="tg-0pky">9,44</td>
<td class="tg-0pky">20200508 06:03:29</td>
</tr>
</tbody>
</table>
I have tried to do something like this:
I tried something like:
SELECT avg(ISNULL(Value,0)), feature FROM V_Value GROUP BY (SELECT avg(Value) from V_Value t1, V_Value t2 where DATEDIFF(minute,t1.DateMis+t1.HourMis,t2.DateMis+t2.HourMis) <= '40' ) order by feature
but it don't work. returns an error on the group by clause
groupings should be:
because every 40 minutes he groups the values of the same feature
the result should be:
Diameter | 59,78
Diameter | 59,924
Weight | 9,088
Weight | 9,562

Answer 1

I took a while to solve this but after some googling and trying stuff I reached this solution.
It rounds the datime down to the desired precision (in this case 40min), and used this rounded time as a grouping column.
SELECT
[rounded_time]
,[feature]
,AVG([value]) AS [value]
FROM (
SELECT
[time]
,DATEADD(mi, DATEDIFF(mi, 0, [time])/40*40, 0) as [rounded_time]
,[feature]
,[value]
FROM V_Value
) as x
To explain what it does you can see the following code sample, it will explain it better than word
DECLARE #SampleData TABLE
(TimeCol Datetime
)
INSERT INTO #SampleData VALUES
('2020-06-04 17:00:00'),('2020-06-04 17:05:00'),('2020-06-04 17:10:00')
,('2020-06-04 17:15:00'),('2020-06-04 17:20:00'),('2020-06-04 17:25:00')
,('2020-06-04 17:30:00'),('2020-06-04 17:35:00'),('2020-06-04 17:40:00')
,('2020-06-04 17:45:00'),('2020-06-04 17:50:00'),('2020-06-04 17:55:00')
,('2020-06-04 18:00:00'),('2020-06-04 18:05:00'),('2020-06-04 18:10:00')
,('2020-06-04 18:15:00'),('2020-06-04 18:20:00'),('2020-06-04 18:25:00')
,('2020-06-04 18:30:00'),('2020-06-04 18:35:00'),('2020-06-04 18:40:00')
SELECT
[TimeCol] AS [Source_Time]
,DATEDIFF(mi, 0, [TimeCol]) AS [MinSince1900] --minutes since 1900-01-01 00:00
,DATEDIFF(mi, 0, [TimeCol])/40*40 AS [MinSince1900_Rounded] --integer divison truncates the decimal part, therefore the result is always rounded down (in this case in blocks of 40min)
,DATEADD(mi, DATEDIFF(mi, 0, [TimeCol])/40*40, 0) AS [Rounded_Time]--add the above minutes back to 1900-01-01 00:00 and you get your grounded datetime
FROM #SampleData