MathGroup Archive 2009

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Re: Graph issue

  • To: mathgroup at
  • Subject: [mg102020] Re: [mg101982] Re: Graph issue
  • From: stu-ultimo <stu-ultimo at>
  • Date: Sun, 26 Jul 2009 05:53:38 -0400 (EDT)
  • References: <h4c1il$h18$> <>

Hi Mark, I guess I had become focused on Combinatorica. Many thanks for  
your solution using GraphUtilities! Stuart

On Sat, 25 Jul 2009 18:16:58 +1000, mark mcclure <mcmcclur314 at>  

> On Jul 24, 6:14 am, "Stuart Nettleton"
> <stuart.nettle... at> wrote:
>> Hi friends, a change appears to have occurred in the behaviour of
>> MakeGraph between Versions 6 and 7.01 (which has caused me some grief!)
>> Below is a small example problem that produces the wrong answer in
>> Version 7.01.0, both 32bit and 64bit. A small topological sort I have
>> included shows Combinatorica's topological sort also gives the wrong
>> answer.
> The problem is not in the TopologicalSort function, but rather the
> MakeGraph function.  Note that the V7 MakeGraph returns a graph
> containing the edge {4,8} (or {h,g}, in the original notation), while
> the V6 version returns the correct {8.4}.  After a quick look at the
> Combinatorica package, I don't really see why.
> Why don't you represent you graph as a list of edge rules and use
> ToCombinatoricaGraph from the GraphUtilities package, if you need to
> use Combinatorica.  Thus,
> Needs["GraphUtilities`"];
> G = Rule @@@ {{g, h}, {a, c}, {b, d},
>   {c, e}, {d, f}, {e, g}, {f, g}};
> GraphPlot[G, DirectedEdges -> True,
>   VertexLabeling -> True,
>   Method -> "CircularEmbedding"]
> It's easy to see that we get the correct graph.  Now, the
> TopologicalSort.
> VertexList[G][[TopologicalSort[
>   ToCombinatoricaGraph[G]]]]
> Mark McClure

Using Opera's revolutionary e-mail client:

  • Prev by Date: Re: Re: Graph issue
  • Next by Date: Re: list of arbitrary length n from user input
  • Previous by thread: Re: Graph issue
  • Next by thread: models in 3d surfaces in Mathematica for fish