SQL Finding similarities - mysql

Question

"Determine if there are actors with the same first name who appeared in the same movie."
This is my task and I'm supposed to do that with subqueries and I just dont really know what else to do. I tried everything with group by, order by and having count but I just don't get to the point where I get the same first name actors with the same movie.
Maybe someone can help me? I am using Sakila Database
SELECT
a.first_name
,(a.last_name)
,a.actor_id
, f.title
FROM actor a
JOIN film_actor fa ON fa.actor_id = a.actor_id
JOIN film f ON f.film_id = fa.film_id
JOIN(SELECT b.first_name, COUNT(*)
FROM actor B
GROUP BY b.first_name
HAVING COUNT(*) > 1 ) b
ON a.first_name = b.first_name
GROUP BY a.last_name
HAVING COUNT(f.title) > 1
ORDER BY a.first_name

Answer 1

You can do this with joins only:
select f.title, a1.first_name, a1.last_name as last_name_1, a2.last_name as last_name_2
from film f
inner join film_actor fa1 on fa1.film_id = f.film_id
inner join film_actor fa2 on fa2.film_id = f.film_id
inner join actor a1 on a1.actor_id = fa.actor_id
inner join actor a2 on a2.actor_id = fa.actor_id
where a1.first_name = a2.first_name and a1.actor_id < a2.actor_id
Starting from the film table, this follows the relationships to actor through film_actor twice, and then filters on diffrerent actors that have the same first name.
As a result, you get tuples of actors that have the same last name and played in the same film. The inequality condition ensures that there are no "mirror" records (that is, each tuple appears only once per film).

Answer 2

I would simply use aggregation:
SELECT fa.film_id, a.first_name,
GROUP_CONCAT(a.last_name) as last_names,
GROUP_CONCAT(a.actor_id) as actor_ids
FROM actor a JOIN
film_actor fa
ON fa.actor_id = a.actor_id
GROUP BY fa.film_id, a.first_name
HAVING COUNT(*) > 1;
Your question doesn't specify what the result set should look like. This returns one row per actors with the same first name in a film. The last names are concatenated into a string as are the actor ids.

Answer 3

How about this:
SELECT f.title, f.film_id, a.first_name, a.last_name, a.actor_id
FROM actor a
JOIN film_actor fa ON fa.actor_id = a.actor_id
JOIN film f ON f.film_id = fa.film_id
WHERE a.first_name IN (
SELECT a2.first_name
FROM actor a2
JOIN film_actor fa2 ON fa2.actor_id = a2.actor_id
JOIN film f2 ON f2.film_id = fa2.film_id
WHERE a2.actor_id <> a.actor_id AND f2.film_id = f.film_id
)
ORDER BY f.title ASC, a.last_name ASC, a.first_name ASC
Explaining query step-by-step
SELECT the needed fields from the joined tables
JOIN the necessary tables
WHERE (here is the subquery) a.first_name is in the set of:
firstnames of actors, different than the current actor (a2.actor_id <> a.actor_id) and the film is the same (f2.film_id = f.film_id)
The subquery in where is a similar select with joins query as the parent query.
PS:
One can do variations on this basic query template:
Eg film_id can be given as parameter, so one can find all actors with same name for specific film.
Also one can group and count how many actors appeared in same film wih same name eg by grouping on film_id and counting.
One can even optimise a bit the query by removing unnecessary joins (eg film.title may not be needed at all) and so on..
The advantage of having single results returned (instead of tuples or aggregates) is that number of actors with same name in same film is not fixed and manipulating the results, eg by grouping and counting or getting further info for each actor, is easier.
The price is a slightly more complex and potentialy slower query.