Re: [igraph] simple example, graph.incidence skips directed parameter

[Gábor Csárdi]

2009/6/14 Jose Quesada <address@hidden>:
> -----BEGIN PGP SIGNED MESSAGE-----
[...]
>> this was a bug for dense matrices only. If you use 'sparse=TRUE', then
>> the 'directed' argument is fine. For dense matrices you cannot trust
>> either 'directed' or 'mode'.
>>
>
> graph.incidence has no sparse parameter. Right?

Yes, I mean if the input object is a sparse matrix.

>> But, remember that the 'mode' argument is not fine for sparse matrices.
>>
>
> Is this the second bug I mentioned? or something else?

Yes, it is exactly that.

G.

> Best,
> - -Jose
>
>
>> G.
>>
>> On Sun, Jun 14, 2009 at 1:25 AM, Jose Quesada<address@hidden> wrote:
>>
>> Hi Gabor,
>>
>> Well, this is not a solution at all, since I have a large graph reading
>> from a sparse matrix; using anything dense will result in the following
>> error:
>>
>> CHOLMOD error: tF
>> Error in asMethod(object) :
>>  Cholmod error 'problem too large' at file:../Core/cholmod_dense.c,
>> line 106
>> Error in as.vector(as(x, "matrix"), mode) :
>>  error in evaluating the argument 'x' in selecting a method for
>> function 'as.vector'
>>
>> Still, when I run graph.incidence on a monster network, it seems to at
>> least respect 'directed'.
>>
>>>>> dg = graph.incidence(m, directed=T, mode="out") # BUG
>>>>>
>>>>>
>>>>> summary(dg)
>> Vertices: 3024016
>> Edges: 11463075
>> Directed: TRUE
>> No graph attributes.
>> Vertex attributes: type, name.
>> No edge attributes.
>>
>>
>> Can I trust this?
>>
>> Best,
>> -Jose
>> Gábor Csárdi wrote:
>>>>> Jose, thanks for the report, this was a bug. I have fixed it, and here
>>>>> is a workaround that works with the currently released version:
>>>>>
>>>>> i <- igraph:::graph.incidence.dense(inc, directed=TRUE, mode="out",
>>>>>                                     weighted=NULL, multiple=FALSE)
>>>>>
>>>>> It is not perfect, because in fact the 'mode' parameter will be always
>>>>> 'out', no matter what you set (for unweighted graphs at least), but
>>>>> you can transpose the incidence matrix and then have the same as you
>>>>> would with mode=="in".
>>>>>
>>>>> Thanks again, best,
>>>>> Gabor
>>>>>
>>>>> On Fri, Jun 5, 2009 at 3:00 PM, Jose Quesada<address@hidden> wrote:
>>>>>> Hi Gabor, Tamar, (all),
>>>>>>
>>>>>> In this simple example, graph.incidence skips directed parameter.
>>>>>>
>>>>>> inc <- matrix(sample(0:1, 15, repl=TRUE), 3, 5)
>>>>>> colnames(inc) <- letters[1:5]
>>>>>> rownames(inc) <- LETTERS[1:3]
>>>>>> i <- graph.incidence(inc, directed=T)
>>>>>> summary(i) # it's not a directed graph
>>>>>>
>>>>>> I have no idea why, because it certainly works on other (bigger)
>>>>>> networks I'm using.
>>>>>> Thoughts?
>>>>>>
>>>>>> Best,
>>>>>> -Jose
>>>>>>
>>>>>> --
>>>>>> Jose Quesada, PhD.
>>>>>> Max Planck Institute,
>>>>>> Center for Adaptive Behavior and Cognition -ABC-,
>>>>>> Lentzeallee 94, office 224, 14195 Berlin
>>>>>> http://www.josequesada.name/
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> igraph-help mailing list
>>>>>> address@hidden
>>>>>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>>>>
>>>>>
>>>>>
>>>
>>>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>
>
> - --
> Jose Quesada, PhD.
> Max Planck Institute,
> Center for Adaptive Behavior and Cognition -ABC-,
> Lentzeallee 94, office 224, 14195 Berlin
> http://www.josequesada.name/
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (MingW32)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iQIcBAEBAgAGBQJKNM64AAoJEGMobUVGH+HK2UcP/RRsLYdNQMD1XG5aX25zgfJ2
> F+Uem+twMgMLTagSPbK5WQl2IEq9iWf4ye5O9u0bWPtC0a5X09ZyOqgUGlfNWOgC
> FKh91gbGajqlqbPMqToN6foNuv5/F3ZLW9Dp1p6B1NsrYnxOON/8QxWGGdRGX9qG
> 0m1MAqCm/6QXpQdmNb04D1sYfwlOxi9QKhkUZokcJ3Ir2JIEPxpEdiEX5e52la43
> uE+MOZbxE/dFAKphEZLJ4Ef7eGDS3H1HymvLamz1gh7bQjuTEEUsjTBU3bZKgSuG
> FJGvzcWUY3+LOrDWZCQX/ulzKFptozCuwvBZbO23buxuQKqtll1R7COr8+kSnyPP
> FTy3idVdbdNM8QpPS8ksJuvf3YDuwFzqdPw3grib4hQMGL0VvOYF0mz1Pt65FEEc
> mL2Es8vcUe1yrxxLEVcNid+8aQi5rqNVqOxIvulF3mjxm7EutfNWQ9U9vGERKTh5
> vXN0AegBNYz/NZBXqoYkW7ehdwf92mGl91qZP/Xv3gPlnCRvIhphivPhrLp+rBb1
> 418/b8TKaS5pPp0DMMqLYZpVYfzoxf/ZyQlOZaWmXPaCJXgMF/6t1rDLKLVNDger
> MJjS9Gb9IBpdCcROrcCbrEUS5MrUjFeoS6HOob8PtCD4ZxMMaKe0aPr4OPLYxalK
> RB9836tZopUx0FgvJtrp
> =WjHH
> -----END PGP SIGNATURE-----
>
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>



-- 
Gabor Csardi <address@hidden>     UNIL DGM