CODEWITHSUNDEEP
neo4jcypher
Jack Daniel
Jack Daniel

Reputation: 2611

Ambiguity of a Cypher Match Query operation

I am retrieving nodes the graph using the id. Once I get a node I am checking whether the node id is present in the list of id's I have. So, here I want to analyze the performance of the query in two ways:

i)Match and Retrieve
ii)Retrieve and Match

First Query:

MATCH (a{uid:4}) RETURN a

Second Query:

MATCH (a) WHERE a.uid in [1,2,3,4] RETURN a

In the first query, Am I retrieving the node when it match uid or else am I retrieving a node and then checking the uid is 4 or not, if not then dropping.

In the second query, I am retrieving the node and then checking whether the node id is present in the list, if yes, then I am returning.

Ambiguity in the functionality in the first query. What exactly happens with the first query?

Upvotes: 0

Views: 46

Answers (1)

Frank Pavageau
Frank Pavageau

Reputation: 11705

You seem to imply the way the query is written defines an order between the operations, but it doesn't. Cypher is a declarative language: you say what you want, it takes care of the execution in the most optimal way possible. There is absolutely no difference between

MATCH (a {uid: 4}) RETURN a

and

MATCH (a) WHERE uid = 4 RETURN a

Both queries express your need to find a node by its uid and the engine will do (as written) a full scan of the nodes to find those matching. If you add a label to the query, and you either have an index or a unicity constraint on uid for that label, then the engine will do an index lookup.

MATCH (a) WHERE uid IN [1, 2, 3, 4] RETURN a

will also do a full scan, except the uid will be checked against the list, whereas

CREATE CONSTRAINT ON (n:Node) ASSERT n.uid IS UNIQUE;
MATCH (a:Node) WHERE uid IN [1, 2, 3, 4] RETURN a

will do a multi-value index lookup.

Upvotes: 3

Related Questions