Message-ID: <CABtg=KnTDzE8qeJX-ek6HkkGsG_iF-fQBsKxyHNyLtJiB+-=Kw@mail.gmail.com>
Date: 2012-11-01T04:10:37Z
From: Gábor Csárdi
Subject: possible bug with largest.cliques in igraph_0.6-3
In-Reply-To: <5091e9a4.2a87320a.31b2.fffff1aaSMTPIN_ADDED@mx.google.com>
Indeed this seems to be a bug, if the graph is directed. The
workaround is to convert it to an undirected graph, the clique
computation ignores direction anyway:
library(igraph)
adj <- matrix(1, nrow=11, ncol=11) - diag(11)
g <- graph.adjacency(adj)
largest.cliques(g)
# [[1]]
# [1] 10 8 1
#
# [[2]]
# [1] 9 7 1
#
# [[3]]
# [1] 8 7 1
largest.cliques(as.undirected(g))
# [[1]]
# [1] 1 2 3 4 5 6 7 8 9 10 11
Btw. please do not send igraph bug reports to r-help, either send them
to igraph-help (see http://igraph.sf.net), to me directly, or, the
best for us, report the bug at https://bugs.launchpad.net/igraph.
Thanks.
I have reported a bug for this, you can follow it here, in case you
are interested:
https://bugs.launchpad.net/igraph/+bug/1073800
On Wed, Oct 31, 2012 at 11:15 PM, Christos Hatzis
<christos.hatzis at nuverabio.com> wrote:
> There is a problem with the largest clique computation in the recent version
> of igraph.
[...]