MySQL Query: limit result respecting substring count - mysql

Question

I have a table like this
ROW ID | CONTENT
------------------------------------------------
test1 | foo, foo, foo
test2 | bar, bar
test3 | foo, foo
test4 | foo, foo, foo, foo
What I want to achieve is query that gives me the rows but limiting it respecting the occurrences of a substring.
Some examples could be:
Limit result to 3 "foo" occurrences -> should return test1
Limit result to 4 "foo" occurrences -> should return test1 and test3
Limit result to 100 "foo" occurrences -> should return test1,test3, test4
Limit result to 7 "foo" occurrences -> should also return test1,test3, test4
Is there any way to do this? Thanks in advance!
P.S. : I should have mentioned that the ',' can be any string without a predictable length.

Answer 1

SQL Fiddle
MySQL 5.5.32 Schema Setup:
CREATE TABLE Table1
(`ROW ID` varchar(5), `CONTENT` varchar(18))
;
INSERT INTO Table1
(`ROW ID`, `CONTENT`)
VALUES
('test1', 'foo, foo, foo'),
('test2', 'bar, bar'),
('test3', 'foo, foo'),
('test4', 'foo, foo, foo, foo')
;
Query 1:
SELECT *
FROM Table1
WHERE ((LENGTH(CONTENT) -
LENGTH(REPLACE(CONTENT, ',', ''))) + 1) < 3
AND SUBSTRING(CONTENT,1,LENGTH('FOO')) = 'FOO'
Results:
| ROW ID | CONTENT |
|--------|----------|
| test3 | foo, foo |
EDIT :
If you are dealing with phrases, it could look like this :
SQL Fiddle
MySQL 5.5.32 Schema Setup:
CREATE TABLE Table1
(`ROW ID` varchar(5), `CONTENT` varchar(48))
;
INSERT INTO Table1
(`ROW ID`, `CONTENT`)
VALUES
('test1', 'foo de foo refe foo'),
('test2', 'bar re bar'),
('test3', 'foo rer ef foo'),
('test4', 'foo rer foo fsdfs foo dfsfe foo')
;
Query 1:
SELECT *
FROM Table1
WHERE (LENGTH(CONCAT(' ',CONTENT,' ')) -
LENGTH(REPLACE(CONCAT(' ',UPPER(CONTENT),' '),
CONCAT(' ','FOO',' '), '')))
/(LENGTH('FOO')+2) < 3 AND
CONCAT(' ',CONTENT,' ') LIKE CONCAT('% ','FOO',' %')
Results:
| ROW ID | CONTENT |
|--------|----------------|
| test3 | foo rer ef foo |

Answer 2

You want to count the number of foos in the list. This is pretty easy:
select t.*
from t
where (char_length(concat(', ', content, ', ')) -
char_length(replace(concat(', ', content, ', '), ', foo, ', '1234567'))
) = 3;
The idea is to replace 'foo' with something that has one fewer character. However, you might want to be careful with 'foobars' and 'barfood' and other strings that could cause a false positive. So, this version just puts the separators at the beginning and end of the string.
Once you have this information, you can do whatever comparisons you would like.

Answer 3

MySQL unfortunately doesn't have any bulit-in function for what you want to do. You need something like SUBSTRING_COUNT, which doesn't exist. What you can do is, based on this answer` calculate that value.
Something like this might work:
SELECT rowid,
(LENGTH(content) - LENGTH(REPLACE(content, 'foo', ''))) / LENGTH('foo') AS cnt
FROM thetable
HAVING cnt > 0 && cnt < 4;
DEMO: http://sqlfiddle.com/#!2/10599/7