Using Date_Sub to subtract year with Aliases column - php

Question

I have two column brand and date_time and during select I assign them to aliases. The query below:
SELECT brand,
YEAR(date_time) AS 'year',
SUM( brand = 'KFC' ) AS kfc
SUM( brand = 'PZH' ) AS pzh,
SUM( brand = 'SUB' ) AS sub
FROM scan_report
GROUP BY YEAR(date_time)
Result to this:
Now here's the confuse part. How can I use Date_Sub function to display data only for the last 5 years. I mean user will enter 2014 and will only display/select record from the last 5 year that means from 2010 to 2014.
Using WHERE such as this gave me empty records:
SELECT
YEAR(date_time) AS 'year',
SUM( brand = 'KFC' ) AS kfc
SUM( brand = 'PZH' ) AS pzh,
SUM( brand = 'SUB' ) AS sub
FROM scan_report WHERE DATE_SUB(2014, INTERVAL 5 YEAR)
GROUP BY YEAR(date_time)
or this
FROM scan_report WHERE DATE_SUB(year = 2014, INTERVAL 5 YEAR)
Any idea how I can retrieve record from the last 5 years of user input? I prefer to solve this problem with mysql, if no solution is found than I'll go with php. Thanks in advance.

Answer 1

Do this simple math in PHP:
$year = 2014; # from $_POST or anywhere else
$from = ($year - 4) . '-01-01 00:00:00';
$to = $year . '-12-31 23:59:59';
And then use this 2 variables in query.
$sql = "
...
WHERE date_time BETWEEN '$from' AND '$to'
...
";
demo
UPDATE (calculation in SQL):
If you don't want to calculate in PHP, you can then calculate in MySQL:
$sql = "
...
WHERE
date_time BETWEEN CONCAT({$year}-4, '-01-01 00:00:00') AND CONCAT({$year}, '-12-31 23:59:59')
...
";
UPDATE (don't care about speed or indexes):
$sql = "
...
WHERE
YEAR(date_time) AND {$year}-4 AND {$year}
...
";

Answer 2

Your problem is that the predicate in the WHERE clause is returning NULL for every row, so no rows are being returned.
The DATE_SUB function takes a DATE expression as the first argument, not an integer. That is, you could specify '2014-01-01' as the first argument, rather than 2014. That would return a non-zero date, which would evaluate to TRUE (in a boolean context) so all rows would be returned.
To include rows that have particular values of date_time, you'd need to use a comparison in the predicate (WHERE clause). To return rows with date_time values of '2009-01-01' and later...
SELECT YEAR(t.date_time) AS year
, SUM( t.brand = 'KFC' ) AS kfc
, SUM( t.brand = 'PZH' ) AS pzh
, SUM( t.brand = 'SUB' ) AS sub
FROM scan_report t
WHERE t.date_time >= DATE_SUB('2014-01-01', INTERVAL 5 YEAR)
GROUP BY YEAR(date_time)
Some additional notes: if you need (or want) to escape an identifier (e.g. a column name, because it's a reserved word), enclose the identifier in backtick characters, not single quotes, e.g.
SELECT YEAR(`t`.`date_time`) AS `year`