Is there some inherent filter in Neo4J Cypher?

0
votes

In the tutorial, there is the following Cypher query:

MATCH (gene:Person)-[:ACTED_IN]->()<-[:ACTED_IN]-(other)
WHERE gene.name="Gene Hackman"
RETURN DISTINCT other;

I generally understand the query. Match all Person nodes that ACTED_IN "any" node, where the gene:Person node has a name property equal to "Gene Hackman".

Then Return a distinct set of "other" nodes that ACTED_IN the same set of "any" nodes as Gene Hackman.

What is returned is a set of nodes, not including the Gene Hackman node.

It seems like the set of nodes being returned would include the Gene Hackman node. There is nothing explicitly filtering out the node containing the name "Gene Hackman" in the Cypher query. So the set of nodes that acted in the same movies as Gene Hackman should include Gene Hackman as well.

I'm assuming there is some inherent rule that is being applied, but the tutorial up to this point hasn't stated what that rule is or I just missed it :^).

Thanks, Todd

2
neo4jcypher

2 Answers

2
votes

The rule is that relationships in a pattern aren't reused. Once you follow the first :ACTED_IN relationship, you can't follow it again in the same pattern match.

0
votes

Wes's answer is ok, and helped me too. In addition please note that this rule holds even if you write the query using 2 paths as 

MATCH (gene:Person)-[:ACTED_IN]->(m),
      (m)<-[:ACTED_IN]-(other)
WHERE gene.name="Gene Hackman"
RETURN DISTINCT other;