Reputation: 3628
I have a MySQL table structured like so:
timestamp | messageVals
-----------------------
1325825203| 46.5
1325825306| 48
1325825404| 43.75
1325825835| 49.625
1325826003| 49.625
1325826055| 47.125
1325794709| 42.875
And so on and so fourth, it has just over 2,000 total rows. I'd like to calculate the average of a set of twenty at a time, and would assumbbly add another column for the average. So every twentieth row would have a value in the avg column that would be an average of the twenty preceding values, which would then reset and do the average of the next twenty.
timestamp | messageVals | avg
--------------------------------
1325825203| 46.5 |
1325825306| 48 |
1325825404| 43.75 |
1325825835| 49.625 |
1325826003| 49.625 |
1325826055| 47.125 |
(20th) 1325794709| 42.875 | 47.15
What's the most easy to understand query that will fill the third column with the averages?
Upvotes: 0
Views: 1638
Reputation: 65587
It's not clear to me whether you actually want to persist the average values in your table or just calculate them in a SELECT query.
Assuming it's the latter, you could do something like this:
set @rownum := 0;
set @sum := 0;
select ts,messageVals,the_avg
from (
select ts,messageVals,
@rownum := (@rownum + 1) as rownum,
@sum := IF(@rownum mod 20 = 1,0 + messageVals,@sum + messageVals) as running_sum,
IF(@rownum mod 20 = 0,@sum / 20,NULL) as the_avg
from so9571582
order by ts
) s;
Upvotes: 2