Neo4j入门（四）批量更新节点属性

本是古典何须时尚 2022-09-11 15:25 564阅读 0赞

本文将用于介绍如何使用py2neo来实现Neo4j的批量更新节点属性。

### 单节点属性更新 ###

首先我们先来看单个节点如何实现节点的属性，比如现有Neo4j图数据库中存在节点，其label为`Test`，属性有`name=上海`，我们需要为该节点增加属性`enname=shanghai`。  
  以下为Python示例代码：

# -*- coding: utf-8 -*-
    from py2neo import Graph
    from py2neo import NodeMatcher
    
    # 连接Neo4j
    url = "http://localhost:7474"
    username = "neo4j"
    password = "******"
    graph = Graph(url, auth=(username, password))
    print("neo4j info: {}".format(str(graph)))
    
    # 查询节点
    node_matcher = NodeMatcher(graph)
    node = node_matcher.match('Test', name="上海").first()
    # 新增enname属性
    node["enname"] = "shanghai"
    graph.push(node)

此时Neo4j图数据库中的该节点属性已经更新，见下图：  
![bb0700b0af404aedb8d8f4ce1c068ab0.png][]

### 多节点单次更新 ###

我们先利用Python代码，在Neo4j图数据库中创建10000个节点，代码如下：

# -*- coding: utf-8 -*-
    from py2neo import Graph, Node, Subgraph
    
    # 连接Neo4j
    url = "http://localhost:7474"
    username = "neo4j"
    password = "******"
    graph = Graph(url, auth=(username, password))
    print("neo4j info: {}".format(str(graph)))
    
    # 创建10000个节点
    node_list = [Node("Test", name=f"上海_{ i}") for i in range(10000)]
    graph.create(Subgraph(nodes=node_list))

如下图所示：  
![创建的10000个节点中的部分节点，节点仅有name属性][10000_name]  
  我们对每个节点进行查询，再一个个为其新增`no`属性，记录下10000个节点的属性更新的耗时。Python代码如下：

# -*- coding: utf-8 -*-
    import time
    from py2neo import Graph, NodeMatcher
    
    # 连接Neo4j
    url = "http://localhost:7474"
    username = "neo4j"
    password = "******"
    graph = Graph(url, auth=(username, password))
    print("neo4j info: {}".format(str(graph)))
    
    # 查询节点
    node_list = []
    for i in range(10000):
        node_matcher = NodeMatcher(graph)
        node = node_matcher.match('Test', name=f"上海_{ i}").first()
        node_list.append(node)
    print("find nodes.")
    
    # 一个个为节点新增no数据
    s_time = time.time()
    for i, node in enumerate(node_list):
        node["no"] = i * 2
        graph.push(node)
    e_time = time.time()
    print("cost time: {}".format(e_time - s_time))

运行结果如下：

cost time: 55.00530457496643

### 多节点批量更新 ###

接下来，我们将节点属性进行批量更新。

# -*- coding: utf-8 -*-
    import time
    from py2neo import Graph, NodeMatcher, Subgraph
    
    # 连接Neo4j
    url = "http://localhost:7474"
    username = "neo4j"
    password = "******"
    graph = Graph(url, auth=(username, password))
    print("neo4j info: {}".format(str(graph)))
    
    # 查询节点
    node_list = []
    for i in range(10000):
        node_matcher = NodeMatcher(graph)
        node = node_matcher.match('Test', name=f"上海_{ i}").first()
        node_list.append(node)
    print("find nodes.")
    print(len(node_list))
    
    
    # 列表划分
    def chunks(lst, n):
        for i in range(0, len(lst), n):
            yield lst[i:i + n]
    
    
    # 批量更新
    s_time = time.time()
    for i, node in enumerate(node_list):
        node["no"] = i * 2
    for _ in chunks(node_list, 1000):
        graph.push(Subgraph(nodes=_))
    e_time = time.time()
    print("cost time: {}".format(e_time - s_time))

运行结果是：

cost time: 15.243977546691895

![多节点批量属性更新后的结果，每个节点新增了no属性][no]

可以看到，批量更新节点属性效率会高很多！当然，如果需要更新的批数比较多，还可以借助多线程或多进程来进一步提升性能。

[bb0700b0af404aedb8d8f4ce1c068ab0.png]: /images/20220828/95eeae91acce4410881ff49e08baf470.png
[10000_name]: /images/20220828/6faf3073a8fd4170af5bd20f7dda38f7.png
[no]: /images/20220828/f0a93cfd8ede4beea75c94558cbf083a.png