SQL select 从一部电影中参与电影次数较多的演员

Question

I'm new to SQL and I'm hardstuck on this query so I would really appreciate some help/advice:

Tables look like

person: id_person name birth

movie: id_movie movie.title

participation: id_person id_movie

And I'm trying to get the id and name of the film and the name of the actor with most participations of that film (in case that 2 got the same ordered by age/alphabetically) for each film in movie.

So far I got:

SELECT 
  DISTINCT movie.id, movie.title, t.name, t.num_movies 
from 
 movie
   JOIN participation on participation.id_movie = movie.id
   JOIN person on participation.id_actor = person.id
   JOIN (
         SELECT 
           person.id as id, person.name as name
           , person.birth, count(movie.id) as num_movies 
        from person 
          JOIN participation on participation.id_actor = person.id 
           JOIN movie on movie.id=participation.id_movie 
        GROUP BY person.id 
        ORDER BY person.birth ASC, person.name ASC
      ) as t on t.id=person.id
ORDER BY t.num_movies DESC

That gives me all the actors with their participations but I just want the ones with most of each film.

Thanks in advance!

Answer 1

A solution would be to use window functions (for MySQL, this requires version 8.0).

You can first join the tables and do a window count of the total appearances of each actor:

select
    m.id movie_id,
    m.title movie_title, 
    p.id person_id,
    p.name person_name,
    count(*) over(partition by p.id) num_movies
from 
    movie m
    inner join participation pm on pm.id_movie = m.id
    inner join person p on p.id = pm.id_person

Then, you can rank the actors of each movie by overall appearances and filter on the top ranked actor of each movie:

select *
from (
    select 
        t.*,
        rank() over(partition by movie_id order by num_movies, person_name desc) rn
    from (
        select
            m.id movie_id,
            m.title movie_title, 
            p.id person_id,
            p.name person_name,
            count(*) over(partition by p.id) num_movies
        from 
            movie m
            inner join participation pm on pm.id_movie = m.id
            inner join person p on p.id = pm.id_person
    ) t
) t     
where rn = 1