Property graph converted using to_networkx function cannot be saved correctly to gexf file

[Jasmine-lxn]

I want to use to_networkx function to convert the data in torch to networkx (retain the attributes of nodes and edges in the graph data), and then save it as a gexf file, but it always prompts an error, what should I do?

[Padarn]

You seem to have posted this twice? It would also usually be easier to help if you provide your code in a code block rather than a screenshot.

In this case the error seems to occur because networkx is assuming that when you have a list as the value object then it is specifying an edge and it's value (from https://github.com/networkx/networkx/blob/7e17ff9007d2e71c2ea2a6d340889c331689343c/networkx/readwrite/gexf.py#L465)

            if isinstance(v, list):
                # dynamic data
                for val, start, end in v:
                    val_type = type(val)

Maybe its a good idea to raise an issue on their github?

[Jasmine-lxn]

Thank you very much for your suggestion, I will try again.