多個可選匹配的優化

Question

在此查詢中，我想匹配所有在其簡歷和我的之間共享節點的用戶，即[：CONTAINS]，或者其中有我感興趣的簡歷中的節點，即[ ：[FOLLOWS] 我想要命令查詢先返回我最常用的查詢（或者：FOLLOWS或：CONTAINS） 此嘗試有效，但代價高昂（而且速度很慢）甚至崩潰了服務器。

您是否看到了更好的方式來獲得想要的用戶？

MATCH (me:User {uuid: "XXX"}) 
 
OPTIONAL MATCH (me)-[:HAS]->(:Resume)-[:CONTAINS]->(n0:Diploma)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u0:User) WHERE me <> u0 WITH me, COLLECT({u: u0, node: n0}) as rows0 
 
OPTIONAL MATCH (me)-[:HAS]->(:Resume)-[:CONTAINS]->(n1:DiplomaMajor)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u1:User) WHERE me <> u1 WITH me, COLLECT({u: u1, node: n1}) as rows1, rows0 
 
OPTIONAL MATCH (me)-[:HAS]->(:Resume)-[:CONTAINS]->(n2:Job)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u2:User) WHERE me <> u2 WITH me, COLLECT({u: u2, node: n2}) as rows2, rows0, rows1 
 
OPTIONAL MATCH (me)-[:HAS]->(:Resume)-[:IN]->(n3:Location)<-[:IN]-(:Resume)<-[:HAS]-(u3:User) WHERE me <> u3 WITH me, COLLECT({u: u3, node: n3}) as rows3, rows0, rows1, rows2 
 
OPTIONAL MATCH (me)-[f:FOLLOWS]->(n4:Location)<-[:IN]-(:Resume)<-[:HAS]-(u4:User) WHERE f.is_wish = 1 WITH me, COLLECT({u: u4, node: n4}) as rows4, rows0, rows1, rows2, rows3 
 
OPTIONAL MATCH (me)-[f:FOLLOWS]->(n5:Diploma)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u5:User) WHERE f.is_wish = 1 WITH me, COLLECT({u: u5, node: n5}) as rows5, rows0, rows1, rows2, rows3, rows4 
 
OPTIONAL MATCH (me)-[f:FOLLOWS]->(n6:DiplomaMajor)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u6:User) WHERE f.is_wish = 1 WITH me, COLLECT({u: u6, node: n6}) as rows6, rows0, rows1, rows2, rows3, rows4, rows5 
 
OPTIONAL MATCH (me)-[f:FOLLOWS]->(n7:Job)<-[:CONTAINS]-(:Resume)<-[:HAS]-(u7:User) WHERE f.is_wish = 1 WITH me, COLLECT({u: u7, node: n7}) as rows7, rows0, rows1, rows2, rows3, rows4, rows5, rows6 
 
UNWIND rows0+rows1+rows2+rows3+rows4+rows5+rows6+rows7 as row WITH DISTINCT row.u AS u, COLLECT(DISTINCT row.node) AS shared_nodes 
 
WHERE LENGTH(shared_nodes) > 0 
 
RETURN u, shared_nodes 
 
ORDER BY LENGTH(shared_nodes) DESC 
 
SKIP 0 LIMIT 10

Answer 1

MATCH (me:User {uuid: "XXX"}) 
     // The path length from one to two, 
     // and the possible relationship types listed 
MATCH (me)-[:HAS|CONTAINS|IN|FOLLOWS*1..2]->(n) 
      // Check for the selected node matching label 
     WHERE ANY(l in ['Diploma','DiplomaMajor','Job','Location'] 
       WHERE l in labels(n)) 
WITH me, n 
     // Find shared user 
MATCH (u)-[:HAS|CONTAINS|IN|FOLLOWS*1..2]->(n) 
     WHERE me <> u 
RETURN u, collect(distinct n) as shared_nodes