For the record, the Python interface of igraph now imposes stricter checks on 
attribute names; only string or Unicode attribute names are now allowed, which 
means that your code would throw a TypeError when trying to construct the graph 
(because csv.DictReader introduces None as an attribute name when it encounters 
a column without a corresponding header name).  

The corresponding changes in the C and Python code are here:

https://github.com/igraph/igraph/commit/cbb82f4412dc22db5f20f82f23fde5e0139fc8cf

--  
T.


On Saturday, 15 February 2014 at 09:40, Tamás Nepusz wrote:

> > Would you let me know when you fix the Graph.DictList() method? I am having 
> > a similar problem with a very large file and I am unable to fix it :-(
>  
>  
>  
> Sure. In the meanwhile, run “del g.es (http://g.es)[None]” as a workaround 
> after you have constructed the graph using Graph.DictList - this gets rid of 
> the edge attribute named None that was mistakenly constructed by 
> Graph.DictList.
>  
> —  
> T.




_______________________________________________
igraph-help mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/igraph-help

Reply via email to