unique genres: Intermediate Cypher Queries

[lorenzofranze]

I wanted to make a variant of the exercise and verify automatically which are the unique genres . This is my solution:

call{

MATCH (n:Movie)

WHERE n.imdbRating IS NOT NULL AND n.poster IS NOT NULL

WITH n {

  .title,

  .imdbRating,

  actors: [ (n)<-[:ACTED_IN]-(p) | p { tmdbId:p.imdbId, .name } ],

  genres: [ (n)-[:IN_GENRE]->(g) | g {.name}]

}

ORDER BY n.imdbRating DESC

LIMIT 4

RETURN collect(n) as listOfFilms

}

call{

  with listOfFilms

  unwind listOfFilms as x

  with uniquegen = [genx{.name} in x.genres where 

    unwind listOfFilms as y

    isEmpty([geny{.name} in y.genres  where x<>y and geny.name=geny.name ]

    )

  ]

return uniquegen

}

return uniquegen

The second call doesn't work since I get an error for the presence of "where " after x.genres in the definition of the list. There I wanted to use a list comprehension to define a list of genres that are unique: that genre is present in only one movie in the list got from the first call. Why I get that error ? 

The problem is in the second call:

With unwind I get x as iterator for listOfFilm then I build the list uniqueGen doing so: I add genx{.name}  in the list (genx is an element of list x) only if there isn't an other element geny with genx = geny and geny got from list y with y iterator on listOfFilm, as previously, I verify also x<>y otherwise the film could be the same.  

[glilienfield]

The error is due to your syntax in the second call statement. There is not an assignment operator in the 'with' clause. Instead, you use the 'as' clause to set a value to a variable. 

I didn't know try to fix your query, as it is really a convoluted approach. You can use the following query if you are just interested in the list of unique genres. 

match(n:Movie)-[:IN_GENRE]->(g) 
with g.name as genre, count(n) as noOfMovies
where noOfMovies = 1
return genre