Check if a node with the same label exists in networkx in python

Question

I am using the following code to build my graph using networkx as follows.

mylist = [('chocolate_pudding', 6.3), ('tim_tam', 178.5), ('bread', 3.5), 
('chocolate_pudding', 8.5), ('tim_tam', 3.0), ('butter', 204.0)]

G = nx.Graph()

for ele in mylist:
    #if node with ele[0] does not exists, add the new node
         G.add_node(ele[0], weight=ele[1])
    #else, update the weight with the new weight value (weight = existing_weight + ele[1])

I want to check if the node with the same label exists before creating a new node. If the node exists I want to update its weight by adding the new value to the existing value.

For example in the above example, I want only 4 nodes chocolate_pudding, tim_tam, bread, butter and their weight should be 14.8, 181.5, 3.5, 204.0 respectively.

I did not find a proper way of doing this in networkx. hence please let me know if this is possible to do in networkx?

Answer 1

You can use ele[0] in G to check whether a label ele[0] is already in there. Updating values are rather straightforward.

import networkx as nx
mylist = [('chocolate_pudding', 6.3), ('tim_tam', 178.5), ('bread', 3.5), ('chocolate_pudding', 8.5), ('tim_tam', 3.0), ('butter', 204.0)]

G = nx.Graph()

for ele in mylist: 
     if ele[0] not in G:
         G.add_node(ele[0], weight=ele[1])
     else:
         new_weight = G.nodes[ele[0]]["weight"] + ele[1]
         # set values
         G.node[ele[0]]["weight"] = new_weight

G.nodes(data=True)        
# NodeDataView({'chocolate_pudding': {'weight': 14.8}, 
#               'tim_tam': {'weight': 181.5}, 
#               'bread': {'weight': 3.5}, 
#               'butter': {'weight': 204.0}})

Edit: This answer is based on networkx 2.0. It seems that before networkx 2.0, nodes is only a function rather than an attribute. Thus, you may need to refer to the documentation for how to update nodes. My guess is G.nodes(data=True)[ele[0]]["weight] = new_weight would work. Or checkout set_node_attributes function.

Answer 2

I think the easiest way is to calculate these values before building the graph

from collections import defaultdict
import networkx as nx

mylist = [('chocolate_pudding', 6.3), ('tim_tam', 178.5), ('bread', 3.5), 
('chocolate_pudding', 8.5), ('tim_tam', 3.0), ('butter', 204.0)]

d = defaultdict(int)

for n, w in mylist:
    d[n] += w

G = nx.Graph()

for n, w in d.items():
    G.add_node(n, weight=w)

print(nx.get_node_attributes(G, 'weight'))
# {'chocolate_pudding': 14.8, 'tim_tam': 181.5, 'bread': 3.5, 'butter': 204.0}