SQL - Select rows where one element appears twice - mysql

Question

I am working on an assignment, and I need to find movies that have been directed by directors that directed more than one movie starring Angelina Jolie. Currently, I have this:
SELECT DISTINCT t.title, n.name
FROM (
SELECT DISTINCT t.id theMovies
FROM name n
INNER JOIN cast_info c
ON (c.person_id = n.id)
INNER JOIN title t
ON (t.id = c.movie_id)
WHERE n.name = 'Jolie, Angelina'
) as newTable
INNER JOIN title t
ON (t.id = theMovies)
INNER JOIN cast_info c
ON (c.movie_id = t.id)
INNER JOIN name n
ON (n.id = c.person_id)
CROSS JOIN role_type
WHERE role = 'director';
What this query currently does is find a list of movies starring Angelina Jolie, and then it lists the directors of those movies. All I need to do now is keep only the rows where the director is present in at least one other row. Any tips?
For reference, here is a diagram of the database I'm using:
http://i.imgur.com/kj8qVgF.png
I'm also rather new to SQL so any suggestions to improve my query would be much appreciated!

Answer 1

I would break this up into several pieces and build up to your final query. If you are new to
SQL, it's good practice to break things into bits and put them back together. With that, I'll restate the goal: find movies that have been directed by directors who have directed a movie with Angelina Jolie.
I would start by getting all movies with Angelina Jolie:
SELECT t.id
FROM name n
JOIN cast_info c ON c.person_id = n.id
JOIN title t ON t.id = c.movie_id
WHERE n.name = 'Jolie, Angelina';
Now, let's get the directors of those movies:
SELECT c.person_id
FROM cast_info c
JOIN title t ON t.id = c.movie_id
JOIN role_type r ON r.id = c.role_id
WHERE r.role = 'director' AND t.id IN(SELECT t.id
FROM name n
JOIN cast_info c ON c.person_id = n.id
JOIN title t ON t.id = c.movie_id
WHERE n.name = 'Jolie, Angelina');
We can modify the above query to group by person_id, having a count(*) greater than one (meaning more than one movie).
SELECT c.person_id
FROM cast_info c
JOIN title t ON t.id = c.movie_id
JOIN role_type r ON r.id = c.role_id
WHERE r.role = 'director' AND t.id IN(SELECT t.id
FROM name n
JOIN cast_info c ON c.person_id = n.id
JOIN title t ON t.id = c.movie_id
WHERE n.name = 'Jolie, Angelina')
GROUP BY person_id
HAVING COUNT(*) > 1;
Now, we need to find movies directed by those directors, and filter so that we don't include movies with Angelina Jolie.
SELECT t.id
FROM title t
JOIN cast_info c ON c.movie_id = t.id
JOIN role_type r ON r.id = c.role_id
WHERE r.role = 'director'
AND c.person_id IN (SELECT c.person_id
FROM cast_info c
JOIN title t ON t.id = c.movie_id
JOIN role_type r ON r.id = c.role_id
WHERE r.role = 'director' AND t.id IN(SELECT t.id
FROM name n
JOIN cast_info c ON c.person_id = n.id
JOIN title t ON t.id = c.movie_id
WHERE n.name = 'Jolie, Angelina')
GROUP BY person_id
HAVING COUNT(*) > 1)
AND t.id NOT IN(SELECT t.id
FROM name n
JOIN cast_info c ON c.person_id = n.id
JOIN title t ON t.id = c.movie_id
WHERE n.name = 'Jolie, Angelina');
I can't test via SQL Fiddle because it is not working at the moment but I will do so as soon as I can. Some stuff might need to be tweaked, but let me know if this helps.

Answer 2

Please try:
SELECT
t.title
, n.name
FROM title t
INNER JOIN cast_info c
ON t.id = c.movie_id
INNER JOIN name n
ON c.person_id = n.id
INNER JOIN role_type r
ON c.person_role_id = r.id
INNER JOIN (
SELECT
c.person_id
, r.id
FROM cast_info c
INNER JOIN role_type r
ON c.person_role_id = r.id
WHERE r.role = 'director'
AND c.movie_id IN (
SELECT DISTINCT
c.movie_id
FROM name n
INNER JOIN cast_info c
ON c.person_id = n.id
WHERE n.name = 'Jolie, Angelina'
)
GROUP BY
c.person_id
, r.id
HAVING COUNT(*) > 1
) d
ON c.person_id = d.person_id
AND r.id = d.id
;
Try a multi-part query "in parts" if it doesn't seem to work, this helps identify where it may be failing
-- 1
SELECT DISTINCT
c.movie_id
FROM name n
INNER JOIN cast_info c
ON c.person_id = n.id
WHERE n.name = 'Jolie, Angelina'
;
-- 2
SELECT
c.person_id
, r.id
FROM cast_info c
INNER JOIN role_type r
ON c.person_role_id = r.id
WHERE r.role = 'director'
GROUP BY
c.person_id
, r.id
HAVING COUNT(*) > 1
;
-- 3
SELECT
c.person_id
, r.id
FROM cast_info c
INNER JOIN role_type r
ON c.person_role_id = r.id
WHERE r.role = 'director'
AND c.movie_id IN (
SELECT DISTINCT
c.movie_id
FROM name n
INNER JOIN cast_info c
ON c.person_id = n.id
WHERE n.name = 'Jolie, Angelina'
)
GROUP BY
c.person_id
, r.id
HAVING COUNT(*) > 1
;