SQL计算集合中出现的平均值

Question

I have a dataset for a school project that consists of a set of movies and genres they belong to. a movie can belong to more than one genre and each is one row in the table (id is pk). eg a small sample

1  Taken  Action
2  Sherlock Holmes Mystery
3  Sherlock Holmes Action
4  Predator        Horror
5  Predator        Action
6  Omen            Horror
7  Pink Panther    Comedy

How would I find that for a given Genre, what is the average number of Genres that a movie in that genre is in. eg on an average a movie in Horror Genre is in 1.5 Genres.

Generally I am used to doing averages on items like salary, numbers but this is slightly different

Answer 1

Please try the following...

SELECT genreListGenre,
       AVG( movieGenreCount ) AS genreMean
FROM ( SELECT genreList.genre AS genreListGenre,
              genreList.movieID AS genreListMovie,
              COUNT( moreGenres.genre ) AS movieGenreCount
       FROM movies AS genreList
       JOIN movies AS moreGenres ON genreList.movieName = moreGenres.movieName
       GROUP BY genreList.genre,
                genreList.movieID
     ) AS genreListGenerator
GROUP BY genreListGenre
ORDER BY genreListGenre;

The inner query joins one instance of movies to another of the same table. So that the fields from each may still be referred to without confusion, each instance is given an alias, genreList and moreGenres . The purpose of the join is to develop a list comprising each genre , each movieID associated with that genre , and each genre associated with that movieID 's corresponding movieName . The first two fields can be determined from one instance of movies , and the third can be drawn from the second instance and is dependent on the value of movieName from the first instance, hence the join being on the shared value of movieName .

Once the list is formed it is grouped by genreList.genre and genreList.movieID and a count of each related genre ( moreGenres.genre ) for that group is made.

The outer query takes the fields returned by the inner query and groups them by genre and for each genre calculates a mean average of the count of each genre associated with each movie that is associated with that group's genre .

If you have any questions or comments, then please feel free to post a Comment accordingly.

Addendum

This code was tested against an instance of movies created (in MySQL) using the following script...

CREATE TABLE movies
(
    movieID     INT,
    movieName   VARCHAR( 50 ),
    genre       VARCHAR( 20 )
);
INSERT INTO movies ( movieID,
                     movieName,
                     genre )
VALUES ( 1, 'Taken',           'Action'  ),
       ( 2, 'Sherlock Holmes', 'Mystery' ),
       ( 3, 'Sherlock Holmes', 'Action'  ),
       ( 4, 'Predator',        'Horror'  ),
       ( 5, 'Predator',        'Action'  ),
       ( 6, 'Omen',            'Horror'  ),
       ( 7, 'Pink Panther',    'Comedy'  );