CODEWITHSUNDEEP
pythongraphnetworkx
artre
artre

Reputation: 321

why does my networkx graph display no edges between the nodes?

I have two lists. One of them has this data structure: in each row first element is the ID, second is the email adress.

     ['0', '[email protected]']
     ['1','[email protected]']

the Second list is a "who wrote to whom" list, in each row with the first ID-Number being the sender and the second being the recipient

     ['0', '4']
     ['0', '6']
     ['1', '4']

Btw, the brackets are not actually part of the lists. this is my jupyter notebook output. the first list has around 2000 and the second list has 40000 rows. In the code below i add the nodes - in this case the email adresses - from the first list. and then i loop through the second list and use the given IDs to create edges between the nodes.

    first =-1
    second =-1
    for row in idsList:
         g.add_nodes_from(row[1])
    for row in dncList:
        for i in range (len(idsList)):
            if (row[0]==idsList[i][0]):
                first=i;
            elif(row[1]==idsList[i][0]):
                second=i
        g.add_edge(idsList[first][1],idsList[second][1])
    nx.draw_networkx(g,with_labels = False, node_size = 30)
    plt.show()

However, i can't see the edges between the nodes. the above code produces the below graph. printing with edges() gives me the list of all the edges, so no problem there. what am i missing here?

Graph

Upvotes: 8

Views: 11928

Answers (3)

eiji56.y1956
eiji56.y1956

Reputation: 43

I tried the following code almost same to yours:

idsList = [ 
  ['0', '[email protected]'],
  ['1','[email protected]'],
  ['2','[email protected]'],
  ['3','[email protected]'],
  ['4','[email protected]']
  ]
dncList = [
  ['0', '4'],
  ['2', '3'],
  ['1', '3'],
  ['1', '2'],
  ['1', '4']
  ]
g = nx.Graph()
first =-1
second =-1
for row in idsList:
      g.add_nodes_from(row[1])
for row in dncList:
    for i in range (len(idsList)):
        if (row[0]==idsList[i][0]):
            first=i;
        elif(row[1]==idsList[i][0]):
            second=i
    g.add_edge(idsList[first][1],idsList[second][1])
nx.draw_networkx(g,with_labels = True, node_size = 300, node_color='skyblue')
plt.show()

then got:

enter image description here

The solution is basically to use add_node instead of add_nodes_from like this:

g = nx.Graph()
for row in idsList:
  g.add_node(row[0], email=row[1])
for row in dncList:
  g.add_edge(row[0], row[1])
nx.draw_networkx(g,with_labels = True, node_size = 300, node_color='skyblue')
plt.show()

enter image description here

Upvotes: 0

melar
melar

Reputation: 212

Your graph looks currently undirected. You should change this and make it a directed graph unless you are supposed to make it undirected. that should solve it. 

g = nx.DiGraph()

Upvotes: 7

Eric Duminil
Eric Duminil

Reputation: 54223

There might be a problem with your edge generation. sec isn't defined for example, and first or second might be kept from a previous iteration.

If you convert your first list to a dict, you won't need a second loop. Here's a faster way to define your graph:

import networkx as nx
import matplotlib.pyplot as plt

l1 = [ ['0', '[email protected]'],
     ['1','[email protected]'],
     ['2','c'],
     ['3','d'],
     ['4','e']
      ]

l2 = [['0', '4'],
     ['0', '2'],
     ['2', '3'],
     ['1', '3'],
     ['1', '2'],
     ['1', '4']]

addresses = dict(l1)

g = nx.Graph()
for address in addresses.values():
    g.add_node(address)

for i1, i2 in l2:
    g.add_edge(addresses[i1], addresses[i2])

nx.draw_networkx(g,with_labels = False, node_size = 30)
plt.show()

It outputs: enter image description here

Upvotes: 0

Related Questions