PHP SQL SELECT sum(value) WHERE x LIKE '%$month%' (foreach) - php

Question

I have a table:
user_id | month | value
1 | firstname_1_2017 | 5
1 | secondname_1_2017 | 4
1 | firstname_2_2017 | 7
1 | secondname_2_2017 | 8
1 | firstname_3_2017 | 5
1 | secondname_3_2017 | 4
1 | firstname_4_2017 | 7
1 | secondname_4_2017 | 8
I need sum "value" for each month with a foreach.
firstname_1_2017 + secondname_1_2017 = 9
firstname_2_2017 + secondname_2_2017 = 15
I have an array for the months:
$array_months = array( '1','2','3','4','5','6','7','8','9','10','11','12' );
So I wrote (in php):
SELECT sum(value)
FROM table
WHERE user_id = $user_id
AND month LIKE '%_$array_months_2017%'
When I write a foreach, the result is the same sum of all values repeat for each month (Example: Jan 48 points, Feb 48 points, Mar 48 points...).
I'd like this result with a foreach:
You have "9" points in January
You have "15" points in February
etc...

Answer 1

You should think about changing your DB design to separate Month and Year from user's name.
Not a good approach, but a solution will be to extract month year from month string and group it to calculate sum for each month.
SELECT SUBSTR(month, LOCATE('_', month) + 1) AS month_year, SUM(value)
FROM users
WHERE user_id = 1
GROUP BY month_year;
Working Fiddle

Answer 2

You can extract the numbers from your month column values with the SUBSTRING_INDEX() function like this, giving 3_2017.
SUBSTRING_INDEX('firstname_3_2017', '_', -2)
You can then transform that to a date-ish string with CONCAT(), yielding 1_3_2017.
CONCAT('1_',SUBSTRING_INDEX('firstname_3_2017', '_', -2))
Then you can turn that into an actual DATE value with STR_TO_DATE().
STR_TO_DATE(CONCAT('1_',SUBSTRING_INDEX('firstname_3_2017', '_', -2)),'%d_%m_%Y')
Cool, now you have a datatype you can actually work with. You can now use an aggregate query to get good stuff. Leveraging Samir's fiddle: http://sqlfiddle.com/#!9/8aabd/6/0
SELECT
SUM(value) value,
user_id,
STR_TO_DATE(CONCAT('1_',SUBSTRING_INDEX(month, '_', -2)),'%d_%m_%Y') month_begin
FROM table
GROUP BY user_id, STR_TO_DATE(CONCAT('1_',SUBSTRING_INDEX(month, '_', -2)),'%d_%m_%Y')
Summary: simple transformations of your month column can give you usable information.
Pro tip: when designing your tables try not to stuff values with different meanings into single columns. Read about normalization.
Pro tip 2: use actual DATE, TIMESTAMP, or DATETIME values for storing time-based information whenever possible. Date arithmetic and date-based indexing are powerful parts of DBMS servers.