[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## Re: [igraph] An elementary question

**From**: |
Yiorgos Adamopoulos |

**Subject**: |
Re: [igraph] An elementary question |

**Date**: |
Thu, 30 Jun 2011 00:08:34 +0300 |

On Tue, Jun 28, 2011 at 6:12 PM, Moses Boudourides
<address@hidden> wrote:
>* I'm sorry to bother you with such a trivial question but it would save*
>* me some time searching through manuals if you were able to give me*
>* directly the answer.*
With my 1 month experience on igraph, I do not find your question a trivial one.
>* Say I have a graph G and all vertices are associated to a binary*
>* attribute which takes the values 1 and 0, i.e., for each vertex the*
>* attribute is either 1 or 0.*
Since I am not accustomed to igraph's attribute interface, I
constructed an array containing values of 0 and 1. I am indexing the
array using the vertex id.
>* (1) G1 which is what it remains from G if we DELETE ALL LINKS*
>* connecting vertices with attribute 0 (but of course keeping in G1 all*
>* vertices of G, although some of them now might turn out to be*
>* isolated)?*
Here is what I did. I constructed an Erdos-Renyi G(n,p) graph. From
this graph I got its adjacency matrix (m) which I copied to two other
matrices (m1 and m0).
I iterated through the vertices of the graph. If the attribute was 0,
then I zeroed the respective line and column in the new adjacency
matrix.
Some drawbacks in my approach that are seen immediately:
1- Isolated vertices are "lost". They can be restored by either
keeping them in a list, inserting a self-loop or some other hack.
2- It is written in C
3- It does not read a CSV. If the list can recommend a parser that
goes from CSV files to an igraph_t maybe this hack I did can improve.
--
http://gr.linkedin.com/in/yiorgos