Finding percentage difference from different values in the same SQL Table? - mysql

Question

I have a table that tracks total values for months against years in a particular location.
Desired Outcome: I wanted to compare a month's value for the current year against last years value. I then wanted to check for a percentage increase.
e.g. 2014 (January) = 140 - 2013 (January) = 150 * 100 = - 6.67
Table Name- donation_tracker
Thank you in advance.

Answer 1

As I understood, You want to get the percent of increase from last year to current year for the Same month for a Particular location. Use the query.
SELECT D1.month, ROUND((D2.Donation_amount- D1.Donation_amount) * 100 /
D1.Donation_amount, 2)
FROM donation_tracker D1
INNER JOIN donation_tracker D2
ON d1.month = D2.month AND D1.year = D2.year - 1
AND D1.Location_ID = D2.Location_ID;

Answer 2

Let's say you need to compare the immediately-completed twelve months with the twelve months prior to that, month-by-month. I am guessing at your table and column names because, well, I don't know them.
Let's build this from the ground up.
Here's a query that will find the most recent twelve months of donations month by month.
SELECT YEAR(donation_date) AS donation_year,
MONTH(donation_date) AS donation_month,
SUM(donation_amount) AS donation_amount
FROM donations
WHERE donation_date >= LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 13 MONTH
AND donation_date < LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 1 MONTH
GROUP BY YEAR(donation_date), MONTH(donation_date)
That gives you a twelve-row result set like this (when NOW() happens to be in the middle of November 2014):
2013 11 145
2013 12 220
2014 1 123
2014 2 11
...
2014 10 45
The trick is picking the right range of donation_date values.
So, now you need two of those result sets, one for mostly-2014 and one for mostly-2013. The one for mostly-2013 looks very similar. You simply back up one more year like this.
SELECT YEAR(donation_date) AS donation_year,
MONTH(donation_date) AS donation_month,
SUM(donation_amount) AS donation_amount
FROM donations
WHERE donation_date >= LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 25 MONTH
AND donation_date < LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 13 MONTH
GROUP BY YEAR(donation_date), MONTH(donation_date)
This is going to be one of those notorious club-sandwich queries, made of those two basic queries. You join them by month like so, then do the percentage computation in the SELECT clause.
SELECT a.donation_month,
a.donation_amount AS this_year,
b.donation_amount AS last_year,
100.0 * (a.donation_amount - b.donation_amount) / b.donation_amount as pct_increase
FROM (
/* this year's query */
) AS a
JOIN (
/* last year's query */
) AS b ON a.donation_month = b.donation_month
ORDER BY a.donation_year, a.donation_month
Here's the whole club sandwich for your server to chew on. Yummy!
SELECT a.donation_month,
a.donation_amount AS this_year,
b.donation_amount AS last_year,
100.0 * (a.donation_amount - b.donation_amount) / b.donation_amount as pct_increase
FROM (
SELECT YEAR(donation_date) AS donation_year,
MONTH(donation_date) AS donation_month,
SUM(donation_amount) AS donation_amount
FROM donations
WHERE donation_date >= LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 13 MONTH
AND donation_date < LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 1 MONTH
GROUP BY YEAR(donation_date), MONTH(donation_date)
) AS a
JOIN (
SELECT YEAR(donation_date) AS donation_year,
MONTH(donation_date) AS donation_month,
SUM(donation_amount) AS donation_amount
FROM donations
WHERE donation_date >= LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 25 MONTH
AND donation_date < LAST_DAY(NOW()) + INTERVAL 1 DAY - INTERVAL 13 MONTH
GROUP BY YEAR(donation_date), MONTH(donation_date)
) AS b ON a.donation_month = b.donation_month
ORDER BY a.donation_year, a.donation_month
Once you stack up the whole club sandwich, it look complicated. But it's actually a stack of simple subqueries.

Answer 3

This should give you an idea :)
Sample data:
CREATE TABLE t
(`month` varchar(3), `year` int, `amount` int)
;
INSERT INTO t
(`month`, `year`, `amount`)
VALUES
('jan', 2013, 150),
('feb', 2013, 180),
('jan', 2014, 140),
('feb', 2014, 160)
;
Query:
select
t1.month, round((t2.amount - t1.amount) * 100 / t1.amount, 2)
from
t t1
inner join t t2 on t1.month = t2.month and t1.year < t2.year;
Result:
| MONTH | ROUND((T2.AMOUNT - T1.AMOUNT) * 100 / T1.AMOUNT, 2) |
|-------|-----------------------------------------------------|
| jan | -6.67 |
| feb | -11.11 |