Neo4j管道数据

Question

Hi there I am on neo4j and I am having some trouble I have one query where I want to return a the a node (cuisine) with the highest percentage like so

// 1. Find the most_popular_cuisine
MATCH (n:restaurants)
WITH COUNT(n.cuisine) as total
MATCH (r:restaurants)    
RETURN r.cuisine , 100 * count(*)/total as percentage
order by percentage desc
limit 1

I am trying to extend this even further by getting the top result and matching to that to get nodes with just that property like so

WITH COUNT(n.cuisine) as total
MATCH (r:restaurants)    
WITH r.cuisine as cuisine , count(*) as cnt
MATCH (t:restaurants)
WHERE t.cuisine = cuisine AND count(*) = MAX(cnt)
RETURN t

Answer 1

I think you might be better off refactoring your model a little bit such that a :Cuisine is a label and each cuisine has its own node.

(:Restaurant)-[:OFFERS]->(:Cuisine)

or

(:Restaurant)-[:SPECIALIZES_IN]->(:Cuisine)

Then your query can look like this

MATCH (cuisine:Cuisine)
RETURN cuisine, size((cuisine)<-[:OFFERS]-()) AS number_of_restaurants
ORDER BY number_of_restaurants DESC  

Answer 2

I wasn't able to use WITH r.cuisine as cuisine , count(*) as cnt in a WITH rather than a RETURN statement, so I had to resort to a slightly more long-winded approach.

There might be a more optimized way to do this, but this works too,

// Get all unique cuisines in a list
MATCH (n:Restaurants)  
WITH COUNT(n.Cuisine) as total, COLLECT(DISTINCT(n.Cuisine)) as cuisineList

// Go through each cuisine and find the number of restaurants associated with each
UNWIND cuisineList as c
MATCH (r:Restaurants{Cuisine:c})
WITH total, r.Cuisine as c, count(r) as cnt
ORDER BY cnt DESC
WITH COLLECT({Cuisine: c, Count:cnt}) as list

// For the most popular cuisine, find all the restaurants offering it
MATCH (t:Restaurants{Cuisine:list[0].Cuisine}) 
RETURN t