CODEWITHSUNDEEP
neo4jcypher
shree11
shree11

Reputation: 535

Fetching specific paths from neo4j Graph database

We are using Neo4j 2.1.4 Community edition.

We are facing some issues in getting the specific paths in neo4j.

Below is the csv file used.

enter image description here

In graph database we are creating Product, Company,Country and Zipcode nodes along with the relationship type as ‘MyRel’ at each level.In the above data we wanted to differentiate each path ,

that is

Mobile, Google,US,88888 -- as path1

Mobile,Goolge,US -- as path2

Mobile,Goolge -- as path3

That’s why we created one more column called Path in data file and maintaining the Path value as a relatioship property. So whenever someone wants to see the different paths he can query based on Relationship property either 1 or 2 or 3. For eample , whenever we query for the relationship property, we should get Mobile,Google ,US

But whenever I do this , in graph it is creating dummy node for Country and Zipcode. This is due to in 2nd and 3rd row the zip and country values are empty(null).

Query used:

LOAD CSV WITH HEADERS FROM "file:C:\\WorkingFolder\\Neo4j\\EDGE_Graph_POC\\newdata\\trial1.csv " as file
MERGE (p:Product {Name:file.Product})
MERGE (comp:Company {Name:file.Company})
MERGE (c:Country {Name:file.Country})
MERGE (zip:Zipcode{Code:file.Zipcode})
CREATE (p)-[:MyRel{Path:file.Path}]->(comp)-[:MyRel{Path:file.Path}]->(c)-[:MyRel{Path:file.Path}]->(zip)

Resultant graph:

enter image description here

So how can I avoid creating dummy nodes ?

Is there any better alternative option to get the proper path?

Thanks,

Upvotes: 1

Views: 97

Answers (2)

Michael Hunger
Michael Hunger

Reputation: 41706

You can filter them out with 

WHERE file.Country <> '' and file.Zipcode <> ''

and split up your CREATE e.g.

CREATE (p)-[:MyRel{Path:file.Path}]->(comp)
WHERE file.Country <> '' and file.Zipcode <> ''
MERGE (c:Country {Name:file.Country})
MERGE (zip:Zipcode{Code:file.Zipcode})
CREATE (comp)-[:MyRel{Path:file.Path}]->(c)-[:MyRel{Path:file.Path}]->(zip)

Upvotes: 0

Jim Biard
Jim Biard

Reputation: 2272

First, A simple solution is to follow the LOAD CSV query with others that clean up your graph. Run the queries

MATCH (zip:Zipcode { Code : ''})<-[r]-()
DELETE zip, r

and

MATCH (c:Country { Name : ''})<-[r]-()
DELETE c, r

You will then have the graph you desire.

Upvotes: 1

Related Questions