from networkx.lattice.grid_2d_graph get adjacency matrix and graph

Question

I have to create a lattice graph of dimensions mxn.

In networkx I would do the following:

N = 5
M = 4 
G = nx.generators.lattice.grid_2d_graph(N,M, periodic=True)

and I would expect a networkx.graph object as a result.

The problem is that, if I call for instance

G.nodes

It does not print the list of nodes as a vector. For instance, if I, instead of using grid_2d_graph, would have used:

G = nx.erdos_renyi_graph(int(N),0.3)
print G.nodes

I would have got a list of numbers

[0,1,2,3,4...,N]

In case of the lattice graph instead I got:

[(0, 1), (1, 2), (3, 2), (0, 0), (3, 3), (3, 0), (3, 1), (2, 1), (0, 2), (2, 0), (1, 3), (2, 3), (4, 3), (2, 2), (1, 0), (4, 2), (0, 3), (4, 1), (1, 1), (4, 0)]

which is like a Matrix.

I would like to get a vector of nodes (as in the other case) and the adjacency matrix of this graph. What can I do?

Joel Joel · Accepted Answer · 2019-01-22T02:01:59

It is a list of nodes, you just don't recognize it because the names aren't integers. For the 2d lattice, the nodes are named by their coordinate. So (0,1) is a node.

Try it:

import networkx as nx
N = 5
M = 4 
G = nx.generators.lattice.grid_2d_graph(N,M, periodic=True)
list(G.neighbors((0,10)))
> [(1, 1), (0, 0), (0, 2), (4, 1)]
G.degree((0,1))
> 4

[note, it appears you're using networkx version 1.11 or earlier. In 2.x, G.nodes is a "NodeView" rather than a list. If you upgrade, you can convert the "NodeView" into a list using list(G.nodes), like what I did with G.neighbors (also not a list in newer networkx versions)].

If you don't want the nodes to be tuples, it is possible to relabel them using nx.relabel_nodes. Let me know if you'd like help with that.

from networkx.lattice.grid_2d_graph get adjacency matrix and graph

1 Answers