2011/9/20 Tiago de Paula Peixoto <tiago@skewed.de>

Hi Giuseppe,

[cut]

The library includes a 'remove_parallel_edges()' function which does
what you want, and is fast.

thanks for the hint, postponing the check after loading the data speeded up the process by 10 times,

however the remove_parallel_edges() does not work in parallel (at least on my machine) and then it's a bit too slow for my purposes.

I'm trying a different approach in order to speed things up.

If you want to mask the parallel edges temporarily, you can use
filtering, as such:

l = label_parallel_edges(g)
g = GraphView(g, efilt=lambda e: l[e] == 0)

Has label_parallel_edges the same complexity of remove_parallel_edges?

Thanks again for your help,

Giuseppe