使用多个分隔符networkx / pandas从.txt文件中添加具有属性的节点

Question

I have a .txt File that has 46 lines, each line stands for a node in a network and then has LOTS of attributes behind it.

Example Name; 03.01.194, Luzern, (LU), Test, Attribute, Other Attribute, 
Kasdasd Alex; 22.12.1957, in Blabla, (ZH), Bürgerorte, Oeschgen (AG),  Zivilstand, 

I'm not sure how I get networkx to see this as a nodelist, some things I thought about, that maybe could work, but do not at the moment

import pandas as pd
import networkx as nx
nodes = pd.read_csv('final.csv', header=None)
nodes

Problem with the code above is that the attributes are separated by commas, but not the nodes.

Another try, where I wanted to open the file, and add nodes line by line but got stuck on the G.add_node() command

G = nx.Graph()
with open('final.txt') as infile:
    for line in infile:
        G.add_node()

Is one of the two the approach to go for or should I try something different?

Also for further analysis, does networkx offer a possibilty to compare attributes of nodes and if they match, create a weighted edge?

Answer 1

You can achieve this by reading the file specifying the delimiter as ';' so that the first element is the node key and the rest are the attributes. Then split the attributes string with the delimiter ',' and add the returned list as a node attribute. I copied the sample you provided in 'test.txt' file and executed the following code.

G = nx.DiGraph()

csv_F = csv.reader(open("test.txt"),delimiter=';')
for row in csv_F:
    attributes=row[1].split(',')
    G.add_node(row[0], attr = attributes)

Then I printed the nodes and their attributes as follows:

for n in G.nodes():
    print 'Node: '  + str(n)
    print 'Atrributes' + str(G.node[n]['attr'])

Result:

Node: Kasdasd Alex

Atrributes: [' 22.12.1957', ' in Blabla', ' (ZH)', ' B\\xc3\\xbcrgerorte', ' Oeschgen (AG)', ' Zivilstand', '']

Node: Example Name

Atrributes: [' 03.01.194', ' Luzern', ' (LU)', ' Test', ' Attribute', ' Other Attribute', ' ']

As for your question in the end, networkx offers such capabilities and more. Have a look on the tutorial here .