Neo4j技术问询：查找未与对应类型节点关联的节点

Hey there, based on your sample graph data (where a1 is linked to b1, while a2/a3 have no connections to any B nodes, and b2 has no connections to any A nodes), here are the exact Cypher queries to meet your two requirements:

You can use either of these two approaches—both will return a2 and a3 as expected:

• Using WHERE NOT EXISTS (explicit check for missing relationships)

  MATCH (a:A)
  WHERE NOT EXISTS((a)-[:LINKS]->(:B))
  RETURN a.title AS unlinked_A_nodes
  

  This query matches every A node, then filters out any that have a LINKS relationship to a B node. The remaining nodes are exactly the ones with no associations to B.

• Using OPTIONAL MATCH + WHERE (check for missing matches)

  OPTIONAL MATCH (a:A)-[r:LINKS]->(:B)
  WHERE r IS NULL
  RETURN a.title AS unlinked_A_nodes
  

  Here, we try to match a LINKS relationship from each A node to a B node. If no such relationship exists, r will be NULL—we filter for those cases to get our target nodes.

Similarly, we can apply the same logic in reverse to get b2:

• Using WHERE NOT EXISTS

  MATCH (b:B)
  WHERE NOT EXISTS((:A)-[:LINKS]->(b))
  RETURN b.title AS unlinked_B_nodes
  

  This matches every B node, then excludes any that have an incoming LINKS relationship from an A node.

• Using OPTIONAL MATCH + WHERE

  OPTIONAL MATCH (:A)-[r:LINKS]->(b:B)
  WHERE r IS NULL
  RETURN b.title AS unlinked_B_nodes
  

  We attempt to match an incoming LINKS relationship from an A node to each B node. Nodes where no such relationship exists (so r is NULL) are our desired results.

内容的提问来源于stack exchange，提问作者Srinath Ganesh