Neo4J: Query to return nodes grouped by relation attributes?

问题

In my case, nodes are connected with one or more relations of type similar. This relation has two attributes fldName and value. For example:

(x)-[r:similar {fldName:'e-mail', value: 'name@x.y.com' }]->(y)
(x)-[r:similar {fldName:'phone', value: '123-45-67' }]->(y)
(q)-[r:similar {fldName:'e-mail', value: 'other@x.y.com' }]->(p)
(q)-[r:similar {fldName:'phone', value: '891-23-45' }]->(p)
(s)-[r:similar {fldName:'e-mail', value: 'neo@x.y.com' }]->(t)

Questions:

1) What is the query to return the following separate groups of results :

group 1: x,y; number of relations: 2
group 2: q,p; number of relations: 2
group 3: s,t; number of relations: 1

2) Is it possible to write this query without a priori knowledge the values of relation attributes?

回答1:

For what you want to achieve, I suggest you use Labels. You could populate the database as follows:

merge (:x)-[:similar {fldName:'e-mail', value: 'name@x.y.com' }]->(:y)
merge (:x)-[:similar {fldName:'phone', value: '123-45-67' }]->(:y)
merge (:q)-[:similar {fldName:'e-mail', value: 'other@x.y.com' }]->(:p)
merge (:q)-[:similar {fldName:'phone', value: '891-23-45' }]->(:p)
merge (:s)-[:similar {fldName:'e-mail', value: 'neo@x.y.com' }]->(:t)

Here you would create nodes with Labels x, y, q, p, s and t.

You can group the results with the following query:

MATCH (x)-[r:similar]->(y)
RETURN labels(x), labels(y), count(r) as relCount

See the Neo4j console for the result.

来源：https://stackoverflow.com/questions/23205781/neo4j-query-to-return-nodes-grouped-by-relation-attributes