RedisGraph技术咨询：如何导入路径数据并查询最短路径？

1. 导入路径数据 #

将你的路径数据转化为RedisGraph可执行的Cypher语句，通过MERGE操作避免重复创建节点和边：

// 创建所有位置节点
MERGE (e:Location {name: 'Entry'})
MERGE (r:Location {name: 'Reception'})
MERGE (c:Location {name: 'Corner'})
MERGE (ex:Location {name: 'Exit'})
MERGE (g:Location {name: 'Gate'})
MERGE (ll:Location {name: 'LiftLobby'})

// 创建带距离属性的连接边
MERGE (e)-[:CONNECTS {distance: 152}]->(r)
MERGE (r)-[:CONNECTS {distance: 176}]->(c)
MERGE (r)-[:CONNECTS {distance: 153}]->(ex)
MERGE (c)-[:CONNECTS {distance: 41}]->(g)
MERGE (c)-[:CONNECTS {distance: 176}]->(r)
MERGE (g)-[:CONNECTS {distance: 53}]->(ll)
MERGE (g)-[:CONNECTS {distance: 41}]->(c)
MERGE (ll)-[:CONNECTS {distance: 53}]->(g)

关键说明 #

• 所有节点统一标记为Location标签，用name属性存储节点标识
• 边使用CONNECTS关系类型，distance属性存储两点间的距离单位
• MERGE关键字会自动校验节点/边是否已存在，无需担心重复插入问题

2. 查询两点间的最短路径 #

提供两种常用的最短路径查询方式，按需选择：

方式1：使用内置最短路径算法（权重优先） #

MATCH (start:Location {name: 'Entry'}), (end:Location {name: 'LiftLobby'})
CALL algo.shortestPath(start, end, 'CONNECTS', 'distance')
YIELD path, weight
RETURN path, weight AS total_distance

方式2：手动匹配路径并计算最短距离 #

MATCH p = (start:Location {name: 'Entry'})-[*]->(end:Location {name: 'LiftLobby'})
WITH p, reduce(total = 0, edge IN relationships(p) | total + edge.distance) AS total_distance
ORDER BY total_distance ASC
LIMIT 1
RETURN p AS shortest_path, total_distance

关键说明 #

• 修改start和end的name属性值，即可查询任意两点间的最短路径
• 方式1调用RedisGraph内置算法，执行效率更高；方式2适合需要自定义路径过滤逻辑的场景

内容的提问来源于stack exchange，提问作者Mohit H