CODEWITHSUNDEEP
neo4jcypher
Sha
Sha

Reputation: 1181

Neo4j Query Optimization for Relations+Nodes

I have a graph data and this data includes millions of data. I try to figure out query optimization methods in neo4j.

For example, I have a query like:

MATCH ((a1:App{appId:1}) <- [:PAID_BY] - (k:Keyword{countryCode:'US'}) - [:PAID_BY] -> (a2:App{appId:2}))
return distinct k.value
limit 50

For this query optimization, which indexes should I create it? Or is there any optimization way for this query?

Note: In this query, I try to find mutual keywords between two apps.

Upvotes: 1

Views: 46

Answers (1)

jose_bacoy
jose_bacoy

Reputation: 12684

Your query is collecting keywords that are common between app1 and app2. You can create indices on App.appId and Keyword.countryCode. Then use this query.

RETURN apoc.coll.toSet( 
 [ (:App{appId:1}) <- [:PAID_BY] - 
 (k:Keyword{countryCode:'US'}) - 
 [:PAID_BY] -> (:App{appId:2}) | k.value]) as keywords


 Where apoc.col.toSet will ensure that the list have unique values
 [ ] is called comprehension and similar to collect() function
 | or pipe is a way to filter data. So you list will contain keyword.value

We love APOC functions and one liner code, isn't?

Upvotes: 2

Related Questions