CODEWITHSUNDEEP
neo4j
Sergio6Rey
Sergio6Rey

Reputation: 401

How to create node, create relationship and remove atribute when loading from CSV?

I have written 3 queries in order to create the nodes :Comments an their properties as well as the relationships with the nodes :Users and :Posts:

USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM "http://neuromancer.inf.um.es:8080/es.stackoverflow/Comments.csv" AS row
    CREATE(n)
    SET n=row
    WITH n AS node
        CREATE (c:Comment {Id: node.Id, CreationDate: node.CreationDate,
                           Score: node.Score, Text: node.Text,
                           UserId: node.UserId, PostId: node.PostId});

MATCH (c:Comment), (u:User)
WHERE toInt(c.UserId) = toInt(u.Id)
    CREATE (u)-[:AUTHOR_OF]->(c)
    REMOVE c.UserId;

MATCH (c:Comment), (p:Post)
WHERE toInt(c.PostId) = toInt(p.Id)
    CREATE (c)-[:INCLUDED_IN]->(p)
    REMOVE c.PostId;

The first query creates the nodes, while the second and third create the relationships between :Comment and :User and :Comment and :Post. However I want to create one single query that creates all of this, in order to make it more efficient. Is this posible and how can I make it? I could not find a way.

Upvotes: 0

Views: 16

Answers (1)

Luanne
Luanne

Reputation: 19373

First, there's no need to create a temporary node n or temporary id properties. Second, it might be more efficient, depending on the size of your file, to do two passes over the file, one to create comments, and the second to match nodes and create relationships. Best to try/profile and see.

Here's what it would look like with just one pass:

USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM "http://neuromancer.inf.um.es:8080/es.stackoverflow/Comments.csv" AS row
CREATE (c:Comment {Id: row.Id, CreationDate: row.CreationDate,Score: row.Score, Text: row.Text})
WITH c
MATCH (u:User) WHERE u.Id = toInt(row.UserId)
CREATE (u)-[:AUTHOR_OF]->(c)
WITH c
MATCH (p:Post) WHERE p.Id=toInt(c.PostId)
CREATE (c)-[:INCLUDED_IN]->(p)

With two passes, 

USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM "http://neuromancer.inf.um.es:8080/es.stackoverflow/Comments.csv" AS row
CREATE (c:Comment {Id: row.Id, CreationDate: row.CreationDate,Score: row.Score, Text: row.Text})

USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM "http://neuromancer.inf.um.es:8080/es.stackoverflow/Comments.csv" AS row
MATCH (c:Comment {Id: row.Id})
MATCH (u:User) WHERE u.Id = toInt(row.UserId)
MATCH (p:Post) WHERE p.Id=toInt(c.PostId)
CREATE (u)-[:AUTHOR_OF]->(c)
CREATE (c)-[:INCLUDED_IN]->(p)

Upvotes: 1

Related Questions