Reputation: 947
I try to do correlation search in Neo4j. I created a simplified example graph (s. image). The question could be:
Is there any correlation between the city a user lives in, the car type, the factory it was produced in and the issue type? So is there any combination that appears more often than all the others? E.g. Users from Munich with a BMW that was produced in Factory XY are likely to have Quality Issues.
In CYPHER I could write the following code:
MATCH (c:City)<--(u:User)-->(car:Car)-->(f:Factory)
RETURN c.name, u.name, car.brand, f.name, count(*)
ORDER BY count(*) DESC
But if my path should go into both directions from the Car-Node and include the Issue Type, how can I write the code?
Upvotes: 1
Views: 160
Reputation: 7478
Are you searching how to split a pattern like this :
MATCH (c:City)<--(u:User)-->(car:Car),
(car)-->(f:Factory),
(car)-->(it:IssueType)
RETURN c.name, u.name, car.brand, f.name, it.name, count(*)
ORDER BY count(*) DESC
Cheers
Upvotes: 1