CODEWITHSUNDEEP
pythongraph-theorynetworkx
Lee
Lee

Reputation: 31040

Get all edges linked to a given node in a networkx graph

Just wondering if there is convenient networkx function that returns a list of edges connected to a given node (or nodes) (e.g. my_node_name) in a graph (e.g. G).

I can do it this way:

edlist=[]
for ed in G.edges():
    if 'my_node_name' in ed:
        edlist.append(ed)

but expect there might be a better way?

Upvotes: 73

Views: 96374

Answers (3)

Antoine
Antoine

Reputation: 936

For directed multi-graphs, I found the following workaround for out-edges:

import networkx as nx

a = nx.MultiDiGraph()
a.add_edge("a", "b", 0)
a.add_edge("a", "b", 1)
a.add_edge("b", "c", 0)
a.add_edge("b", "c", 1)

# This is too much:
print(a.edges)
# OutMultiEdgeView([('a', 'b', 0), ('a', 'b', 1), ('b', 'c', 0), ('b', 'c', 1)])

# This is insufficient:
print(a.in_edges("b"))
# InMultiEdgeDataView([('a', 'b'), ('a', 'b')])

# This is just right
print(a["a"])
# AdjacencyView({'b': {0: {}, 1: {}}})
print(a["a"]["b"])
# AtlasView({0: {}, 1: {}})

For in-edges, you should probably first use a.in_edges(node) and then a[source][node] to get the edge types.

Once you get all the triplets t = (source, sink, type), you can access the edge by calling a.edges[t].

Upvotes: 1

donkopotamus
donkopotamus

Reputation: 23166

You can use the method edges on a node for an un-directed graph: 

G.edges(['my_node_name'])

or the function edges

networkx.edges(G, ['my_node_name'])

But for directed graphs the above method will only give the out-edges; there you need to call and combine both in_edges() and out_edges().

Upvotes: 18

Joel
Joel

Reputation: 23827

If the graph is undirected, you can use 

G.edges(node)

In networkx 2.x this is an EdgeDataView object. In networkx 1.x this is a list - if you want a generator in 1.x rather than getting the whole list, G.edges_iter(node) works (this no longer exists in 2.x).

If the graph is directed the command above will not give the in-edges. Use 

G.in_edges(node)
G.out_edges(node)

These are views in 2.x. In 1.x these are lists and there are generator options: G.in_edges_iter(node) and G.out_edges_iter(node)

Upvotes: 116

Related Questions