MySQL ONLY IN() equivalent clause

问题

I am giving a very abstract version of my question here, so please bear with me. I have a query that will check whether a particular body has certain multiple parameters of same type. Example, a boy has multiple selection as far as chocolates are concerned. But, I want to choose boys from the table who have exactly the chocolates I mention. Not more not less and not 'LIKE' or not 'IN()'.

SELECT boy_id from boys_chocolates WHERE chocolate_id ONLY IN('$string');

..where of course '$string' is a PHP variable containing comma separated values of only those chocolates I want to use to pull the boys.

I know this is invalid MySQL statement, but is there any valid equivalent to this?

EDIT:

This is more comprehensive query which gets records in special cases, but not always.

SELECT boys.* FROM schools_boys INNER JOIN boys ON boys.boy_id=schools_boys.boy_id
INNER JOIN chocolates_boys a ON a.boy_id=boys.boy_id INNER JOIN schools
ON schools.school_id=schools_boys.school_id WHERE a.chocolate_id IN(1000,1003)
AND 
            EXISTS
            (
                    SELECT 1
                    FROM chocolates_boys b
                    WHERE a.boy_id=b.boy_id
                    GROUP BY boy_id
                    HAVING COUNT(DISTINCT chocolate_id) = '2'
                    )

                GROUP BY schools_boys.boy_id HAVING COUNT(*) = '2'

Boys Table
+--------+-------------+
| id     | boy         |
+--------+-------------+
| 10007  | Boy1        |
| 10008  | Boy2        |
| 10009  | Boy3        |
+--------+-------------+

Chocolates Boys Table
+----+---------+--------------+
| id | chocolate_id | boy_id |
+----+--------------+---------+
| 1  | 1000         | 10007   |
| 2  | 1003         | 10007   |
| 3  | 1006         | 10007   |
| 4  | 1000         | 10009   |
| 5  | 1001         | 10009   |
| 6  | 1005         | 10009   |
+----+--------------+---------+

Nothing happens when I select 1000 alone to pull two boys (or) 1000 and 1003 to pull out the boy with ID 10007.

回答1:

this problem is called Relational Division

SELECT boy_id 
FROM   boys_chocolates 
WHERE  chocolate_id IN ('$string')
GROUP  BY boy_id 
HAVING COUNT(DISTINCT chocolate_id) = ? -- <<== number of chocolates specified

example:

SELECT boy_id 
FROM   boys_chocolates 
WHERE  chocolate_id IN (1,2,3,4)
GROUP  BY boy_id 
HAVING COUNT(DISTINCT chocolate_id) = 4

SQL of Relational Division

however, if the chocolate_id is unique for every boy_id, DISTINCT keyword is optional.

SELECT boy_id 
FROM   boys_chocolates 
WHERE  chocolate_id IN (1,2,3,4)
GROUP  BY boy_id 
HAVING COUNT(*) = 4

UPDATE 1

...I want to choose boys from the table who have exactly the chocolates I mention. Not more not less...

SELECT boy_id 
FROM   boys_chocolates a
WHERE  chocolate_id IN (1,2,3,4) AND
        EXISTS 
        (
            SELECT  1
            FROM    boys_chocolates b
            WHERE   a.boy_ID = b.boy_ID
            GROUP   BY boy_id
            HAVING  COUNT(DISTINCT chocolate_id) = 4
        )
GROUP  BY boy_id
HAVING COUNT(*) = 4

SQLFiddle Demo

来源：https://stackoverflow.com/questions/16461707/mysql-only-in-equivalent-clause

标签

mysql

clause

relational-division