Question

The table:

id | category   (there is a index on id & category)
-----------
1  |  1
1  |  7
1  |  3

2  |  1
2  |  2
2  |  4

3  |  1
3  |  6
3  |  3

SELECT DISTINCT
    category
FROM 
    many_to_many e1     
WHERE id IN 
    (
    SELECT DISTINCT
        e1.id
    FROM 
        many_to_many e1           
    INNER JOIN
        many_to_many x1 
      ON e1.id=x1.id
    WHERE
        e1.category IN (3)
    )

I like to get retured: ** 6, 1, 7** (what I do get with the query above)
It seems to me this query is not gone preform well, because the sub query searches all the ID’s and this list can be huge.

Also it doesn’t matter how many but if the ID is related.
So for performance if there would be 100 id’s checking once for each unique category to be populated would be enough.

Secondly I use an other query (the sub query )the get all id’s who contain the category:

SELECT DISTINCT
        e1.id
    FROM 
        many_to_many e1           
    INNER JOIN
        many_to_many x1 
      ON e1.id=x1.id
    WHERE
        e1.category IN (3)

 returns: 3 & 1

What would be the most efficient way to query the result I like. Is there an more efficient (better preforming) solution?
Should I use one instead of two query’s?

Answer 1

SELECT DISTINCT
        x1.category
    FROM 
        many_to_many e1           
    INNER JOIN
        many_to_many x1 
      ON e1.id = x1.id
    WHERE
        e1.category = 3          --- IN (3) 
      AND
        x1.category <> 3         --- NOT IN (3)

You shouldl check the plan of the query with EXPLAIN. A compound index on (category, id) might be good for this query.