Writing a custom binary format (.gt file) in spark

graph-tool
1

I've got a dataframe inside of Apache spark which I'd like to output as a
.gt file.

The dataframe is a table of edges between vertices like so:
0 9
0 8
9 1

I read the guide on the custom gt file format, but how do I actually write
one for my edges dataframe? That too, inside of spark? Any ideas/examples
you could point me to?

Any help is appreciated, thank you!

attachment.html (602 Bytes)

2

The format is documented here:

    https://graph-tool.skewed.de/static/doc/gt_format.html

But if you have a list of edges, you can save it as a binary numpy array,
and load the graph with Graph.add_edge_list(). It should be of comparable speed.

Best,
Tiago