DFS find_cycle method is outputting a cycle where there is none

[Taylor Do]

Hi,

_________________________________________

#test

import networkx as nx

G=nx.DiGraph()

for x in range(0, 11):

  G.add_node(x)

G.add_edges_from(

[ (0, 36), 

 (0, 38), 

 (1, 36), 

 (1, 39), 

 (2, 37), 

 (18, 0), 

 (19, 0), 

 (19, 39), 

 (20, 1), 

(20, 37), 

 (20, 39), 

 (21, 37), 

 (21, 40), 

 (22, 38), 

 (38, 1), 

 (39, 1), 

 (39, 40),

 (40, 2), 

 (40, 38) ])

print(nx.find_cycle(G, source=0))

**After the code is run: output => [(1, 39), (39, 1)]

______________________________________________

When the code above is run, it outputs a directed cycle of [(1, 39), (39, 1)] which obviously is not correct as the graph does not contain a cycle from 0 (as you can tell from the edges). 

Is there something that I am missing here or is there a problem with the find_cycle method that needs to be patched?

Thanks, 

Taylor Do

[Dan Schult]

Take a look at the docs:

nx.find_cycle uses a depth first traversal starting at source.

If it finds a cycle during that traversal it reports the cycle.

This cycle might not include the source.

[Taylor Do]

Thank you for your response. 

Is there a method from a library that you know of that would return a cycle that contains the source node?

[Dan Schult]

https://networkx.org/documentation/stable/reference/algorithms/cycles.html