I have a table with columns date
and time_spent
. I want to find for each date D the sum of the values of 'time_spent' for the period of time : (D-7 - D), ie. past week + current day.
I can't figure out a way to do this, as I can only find examples for a total sum and not a sum over a variable period of time.
Here is a dataset example :
CREATE TABLE rolling_total
(
date date,
time_spent int
);
INSERT INTO rolling_total VALUES ('2013-09-01','2'),
('2013-09-02','1'),
('2013-09-03','3'),
('2013-09-04','4'),
('2013-09-05','2'),
('2013-09-06','5'),
('2013-09-07','3'),
('2013-09-08','2'),
('2013-09-09','1'),
('2013-09-10','1'),
('2013-09-11','1'),
('2013-09-12','3'),
('2013-09-13','2'),
('2013-09-14','4'),
('2013-09-15','6'),
('2013-09-16','1'),
('2013-09-17','2'),
('2013-09-18','3'),
('2013-09-19','4'),
('2013-09-20','1'),
('2013-09-21','6'),
('2013-09-22','5'),
('2013-09-23','3'),
('2013-09-24','1'),
('2013-09-25','5'),
('2013-09-26','2'),
('2013-09-27','1'),
('2013-09-28','4'),
('2013-09-29','3'),
('2013-09-30','2')
Result would look like :
date | time_spent | rolling_week_total
2013-09-01 | 2 | 2
2013-09-02 | 1 | 3
2013-09-03 | 3 | 6
2013-09-04 | 4 | 10
2013-09-05 | 2 | 12
2013-09-06 | 5 | 17
2013-09-07 | 3 | 20
2013-09-08 | 2 | 22
// now we omit values that are older than seven days
2013-09-09 | 1 | 21
2013-09-10 | 1 | 21
...
And one more solution
SELECT r1.date, r1.time_spent, sum(r2.time_spent) AS rolling_week_total
FROM rolling_total AS r1 JOIN rolling_total AS r2
ON datediff(r1.date, r2.date) BETWEEN 0 AND 7
GROUP BY r1.date
ORDER BY r1.date
LIMIT 8