如何编写Neo4j Cypher查询并合并拥有至少2个公共节点的节点

场景1：查询两两共享至少2个公共节点的节点对 #

适用于仅存在ROW1、ROW2两个节点共享ID_B、ID_C两个公共节点的图谱场景：

// 匹配两个不同的Row节点，及二者共同关联的中间ID节点
MATCH (r1:Row)-[:REL]->(common_id:ID)<-[:REL]-(r2:Row)
WHERE id(r1) < id(r2)  // 避免重复返回r1-r2、r2-r1同一对
WITH r1, r2, collect(DISTINCT common_id) AS shared_ids
WHERE size(shared_ids) >= 2  // 筛选至少2个公共节点的配对
// 返回两个节点详情及公共节点列表
RETURN r1 AS row1_detail, r2 AS row2_detail, shared_ids AS common_nodes

注：请将代码中的:Row（行节点标签）、:REL（行节点与ID节点的关系类型）、:ID（公共ID节点标签）替换为你实际的图谱结构定义。

场景2：多节点连锁共享公共节点的合并查询 #

适用于存在Row1、Row2、Row3两两之间均满足至少2个公共节点的图谱场景，可合并返回所有关联的Row节点：

// 先匹配所有满足两两共享≥2公共节点的Row对
MATCH (r1:Row)-[:REL]->(common_id:ID)<-[:REL]-(r2:Row)
WHERE id(r1) < id(r2)
WITH r1, r2, collect(DISTINCT common_id) AS shared_ids
WHERE size(shared_ids) >= 2
// 收集所有关联的Row节点去重，同时收集所有配对的公共节点
WITH collect(DISTINCT r1) + collect(DISTINCT r2) AS all_row_list, collect({row_pair: [r1.name, r2.name], shared: shared_ids}) AS pair_shared_info
UNWIND all_row_list AS row
WITH DISTINCT row, pair_shared_info
// 合并返回所有相关Row详情及各配对的公共节点信息
RETURN collect(DISTINCT row) AS all_related_row_details, pair_shared_info AS all_pair_common_nodes

如果仅需要返回所有符合条件的Row节点的完整属性，将RETURN语句替换为RETURN collect(DISTINCT row{.*}) AS row_detail_list即可。

内容的提问来源于stack exchange，提问作者SSK