Do not use named path (unless it is required)

Named path in a query always comes at an additional cost, which can add penalties in terms of higher latency and memory usage. Consider the following query:

MATCH p = (n)-[:commentedOn]->(m)
WITH p, m, n, n.score + m.score as total
WHERE total > 100 
MATCH (m)-[:commentedON]->(o)
WITH p, m, n, distinct(o) as o1
RETURN p, m.name, n.name, o1.name

In the above query, assuming we only want to know the properties of the nodes, the use of path “p” is unnecessary. By specifying the named path as a variable, the aggregation operation using DISTINCT will get expensive both in terms of time and memory usage. A more optimized version of above query could be:

MATCH (n)-[:commentedOn]->(m)
WITH m, n, n.score + m.score as total
WHERE total > 100 
MATCH (m)-[:commentedON]->(o)
WITH m, n, distinct(o) as o1
RETURN m.name, n.name, o1.name