Neo4j-关系建模问题

Question

I have 2 CSV files exported from mysql :

# Disease CSV Headers # Disease Master ID (autoincrement, pk) disease_name

# Tests CSV Headers # Test Master ID (autoincrement, pk), test_name, parent_disease_ID (points to ID column in Disease.
Master tbl)

I run following cypher commands :

    LOAD CSV WITH HEADERS FROM.   
   "http://localhost/disease_mstr.csv" AS line
    MERGE (d:Disease {did: toInteger(line.ID),  diseasename: 
    line.disease_name})

    LOAD CSV WITH HEADERS FROM.  
   "http://localhost/test_mstr.csv" AS line
    MERGE (d:Tests {tid: toInteger(line.ID),  testname: 
    line.test_name, did: toInteger(line.parent_disease_ID)})

   MATCH (t:Tests), (d:Disease) CREATE (t:Tests)- 
   [r:TEST_FOR]->(d:Disease) RETURN t, r, d

Above cypher returns one disease connected to many tests whereas i want just the reverse! Can someone please correct me?

Answer 1

You could create the disease nodes, test nodes and relationship between test and disease nodes in one pass of the tests file.

LOAD CSV WITH HEADERS 
FROM "http://localhost/test_mstr.csv" 
AS line
MERGE (disease:Disease {did: toInteger(line.parent_disease_ID)})
MERGE (test:Tests {tid: toInteger(line.ID), testname: 
line.test_name})
MERGE (test)-[r:TEST_FOR]->(did)

And then update the disease names after the fact in a second pass.

LOAD CSV WITH HEADERS 
FROM "http://localhost/disease_mstr.csv" AS line
MERGE (d:Disease {did: toInteger(line.ID)})
SET d.diseasename = line.disease_name

Answer 2

[EDITED]

In the query that creates the relationships, you need to filter for Tests and Disease nodes that share the same did value:

MATCH (t:Tests), (d:Disease)
WHERE t.did = d.did
MERGE (t)-[r:TEST_FOR]->(d)
RETURN t, r, d;

This query also replaced CREATE with MERGE to avoid creating duplicate TEST_FOR relationships between the same t and d pair. If you already have such duplicate relationships, delete them first.

Also, for efficiency, you should consider creating in index on either :Disease(did) or :Tests(did) -- whichever has the most instances.