如何将本地目录的CSV数据导入Memgraph并使用Python运行DFS算法？

一、加载本地CSV到Memgraph #

1. 准备CSV数据 #

确保CSV文件结构清晰，区分节点和边数据：

• 节点文件（如nodes.csv）：包含节点唯一标识、属性和标签

  id,name,label
  1,Alice,Person
  2,Bob,Person
  3,Charlie,Person
  

• 边文件（如edges.csv）：包含起始节点ID、目标节点ID和关系类型

  start_id,end_id,relationship
  1,2,FRIENDS_WITH
  2,3,FRIENDS_WITH
  1,3,WORKS_WITH
  

2. 导入数据到Memgraph #

方式1：直接通过Memgraph Lab/mgconsole执行Cypher命令

连接Memgraph后，执行以下命令导入节点和边：

• 导入节点：

  LOAD CSV FROM "/本地路径/nodes.csv" WITH HEADER AS row
  CREATE (n:Person {id: toInteger(row.id), name: row.name});
  

• 导入边：

  LOAD CSV FROM "/本地路径/edges.csv" WITH HEADER AS row
  MATCH (a:Person {id: toInteger(row.start_id)}), (b:Person {id: toInteger(row.end_id)})
  CREATE (a)-[r:FRIENDS_WITH]->(b);
  

方式2：Docker运行Memgraph时挂载本地目录

如果用Docker启动Memgraph，先挂载CSV所在本地目录到容器：

docker run -p 7687:7687 -v /本地CSV目录:/data/csv memgraph/memgraph

然后在Cypher中使用容器内路径：

LOAD CSV FROM "/data/csv/nodes.csv" WITH HEADER AS row
CREATE (n:Person {id: toInteger(row.id), name: row.name});

二、用Python实现DFS算法 #

1. 安装依赖 #

安装Memgraph的Python客户端mgclient：

pip install mgclient

2. 连接Memgraph并获取图数据 #

import mgclient

# 建立数据库连接
conn = mgclient.connect(host='localhost', port=7687)
cursor = conn.cursor()

# 提取所有节点信息
cursor.execute("MATCH (n:Person) RETURN n.id, n.name")
nodes = {row[0]: row[1] for row in cursor.fetchall()}

# 构建邻接表
cursor.execute("MATCH (a)-[r]->(b) RETURN a.id, b.id")
adjacency = {node_id: [] for node_id in nodes.keys()}
for a_id, b_id in cursor.fetchall():
    adjacency[a_id].append(b_id)

3. 实现DFS遍历 #

方式1：手动实现DFS逻辑

def dfs(start_id, adjacency, visited=None):
    if visited is None:
        visited = set()
    visited.add(start_id)
    print(f"访问节点: {nodes[start_id]}")
    # 遍历所有邻接节点
    for neighbor_id in adjacency[start_id]:
        if neighbor_id not in visited:
            dfs(neighbor_id, adjacency, visited)
    return visited

# 从ID为1的节点开始DFS
dfs(1, adjacency)

方式2：使用Cypher内置路径查询实现DFS

直接通过Cypher查询获取DFS路径：

cursor.execute("""
    MATCH path=(n:Person {id:1})-[*]->(m:Person)
    RETURN nodes(path) AS path_nodes
""")
# 打印所有DFS路径
for row in cursor.fetchall():
    path_names = [node['name'] for node in row[0]]
    print(f"DFS路径: {' -> '.join(path_names)}")

4. 关闭连接 #

conn.close()

内容的提问来源于stack exchange，提问作者MPesi