Python 将有向的流数据合并为无向的流

需求：有一份人口流动数据，是有向的人口流动，尝试合并为合并的流动，数值为双向流动之和。

解决方法：一开始尝试采用merge的方法，最后总是差一点东西，最后得到间接的networkx中的有向图转无向图的方法，再对每个边进行计算，代码如下：

  # 读取流数据为dataframe，主要是source,target,value字段
    input_path = "D:\\wyx2020\\flow\\"
    df = pd.read_csv(input_path+"flow.csv", header=0, encoding='gbk')
    
    # 采用network中的有向图到无向图的转换
    # 构建有向图
    G = nx.from_pandas_edgelist(df, 'source', 'target', edge_attr=True, create_using=nx.DiGraph())
    undirected = G.to_undirected() #转变为无向图
    
# 依然需要遍历每条边，进行属性计算
    for edge in undirected.edges(data=True):
        direction_1 = df.loc[(df['source'] == edge[0]) & (df['target'] == edge[1])]
        direction_2 = df.loc[(df['source'] == edge[1]) & (df['target'] == edge[0])]
        edges = pd.concat([direction_1, direction_2])
        edge[2]['bidirect'] = 1 if (not direction_1.empty) & (not direction_2.empty) else 0
        edge[2]['Value'] = edges['Value'].sum()

# 将得到的dataframe输出
    nx.to_pandas_edgelist(undirected).to_csv(input_path+'undir_flow_del_inter_city.csv',
                                             header=True, index=False, encoding='gbk')

参考资料：https://stackoverflow.com/questions/52255502/how-to-transform-directed-graph-to-undirected-graph-while-recording-information