Re: [igraph] Less vertices in output than input

[Laurens Rietveld]

Hi Gang,

Many thanks for the offer. I've gone for a quick and dirty solution (custom script running on big memory server) in generating the pajek file. It's still running though, so not sure whether the pajek file will save me.

Best,

Laurens

On Fri, May 3, 2013 at 5:00 PM, 四正（红砖） <address@hidden> wrote:

Hi, Laurens,

I have a program made by myself to transform the edgelist file into pajek file.

If the file is large, it buffers data in a MSSQLSERVER database. You need .NET4 to run it. Do you want it?

Gang

------------------ Original ------------------

From:  "Laurens Rietveld"<address@hidden>;

Date:  Fri, May 3, 2013 10:34 PM

To:  "Help for igraph users"<address@hidden>;

Subject:  Re: [igraph] Less vertices in output than input

Hi Gabor,

Thanks for replying. I can provide an example, though it is quite big (~3GB): http://aers.data2semantics.org/input.tsv

I've tried smaller graphs to reproduce this issue with, but couldnt find any.

I will try to get this into pajek format, though it'll be some work. I'd hope to avoid this as much as possible (I have bigger graphs around 130GB as well..).

If I understand correctly, the number of isolates might be an issue here? Thing is, when I take a look at my graph, I'm sure some of the vertices have incoming edges. However, when I calculate the indegree, some of these vertices are still missing..

Best, Laurens

On Fri, May 3, 2013 at 3:47 PM, Gábor Csárdi <address@hidden> wrote:

You need to send us a complete example that reproduces a problem. Most probably you have isolate vertices, and from an edge list file format it is sometimes not possible to know how many isolates you have. Use another file format, like Pajek, that is almost the same as an edge list, but specified the number of vertices explicitly.

Gabor

On Fri, May 3, 2013 at 9:12 AM, Laurens Rietveld <address@hidden> wrote:

I'm running igraph to do some experiments, and for some graphs I notice the number of vertices in the analysis output table (e.g. the output of evcent (graph, directed = FALSE)) is less than the vertices used as input (edge list).

Because this only happens for some typical graphs (quite disconnected), I'm assuming this occurs because there is no weight assigned to this vertice (i.e. value zero). Is this true? Or is something else going wrong here?

Many thanks,

Laurens

_______________________________________________
igraph-help mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/igraph-help

--
Gabor Csardi <address@hidden> MTA KFKI RMKI