Re: gGeoManager->GetCurrentNodeId()

Hi Mark,

As you have noticed, the method will not give you really unique id's - in normal case it is used as a backup method to check if the current tracking location have changed or not. There are few ways to do what you want. First is to call CloseGeometry("di") that will enable really unique node id's. You should be aware that this needs to allocate space in memory as 2*Nnodes words where Nnodes is the total number of nodes reported by CloseGeometry(). In this case you can get the unique id using GetCurrentNodeId() but also make a given id current: gGeoManager->CdNode(id) to get its matrix or whatever. I wouldn't do this however for calorimeters...
The second method is to account at transportation time the current node and copy number (e.g. pad # or read out channel) and parents in the current branch (gGeoManager->GetParent(i)) (e.g. layer/sector/module). The information can be retrieved fast this way, no need to store unique numbers for each channel in different modules...

Best,
Andrei

Mark Messier wrote:
> Hi All,
>
> I'm working with a geometry and need to assign unique read out channel
> numbers to the lowest level nodes in my geometry. I thought I could use
> the "gGeoManager->GetCurrentNodeId()" to map each detector cell uniquely
> to an integer, but when I do that, I discover that my geometry has
> duplicate ID numbers (that is, TGeoNodes located along different paths
> that return the same ID number). Perhaps I am misinterpreting what is
> meant my the NodeId's?
>
> Looking at the code that assigns NodeId's it seems to me that duplicate
> ID's are expected:
>
> Int_t TGeoNodeCache::GetNodeId() const
> {
> // Get unique node id.
> Long_t id=0;
> for (Int_t level=0;level<fLevel+1; level++)
> id += (Long_t)fNodeBranch[level];
> return (Int_t)id;
> }
>
> This sum which defines id's has an obvious degeneracy when
> fNodeBranch[0,1] = 1,2 and fNodeBranch[0,1] = 2,1. That's
> oversimplified, of course, since the fNodeBranch[]'s are pointers, but
> other such degeneracies must exist.
>
> Any tips on how to efficiently map a geometry path to a single ID number?
>
> Thanks,
> Mark
> ========================================================================
> Mark Messier messier_at_indiana.edu
> Department of Physics http://www.physics.indiana.edu/~messier
> Indiana University Phone: (812) 855-0236
> Bloomington, IN 47405 Fax: (812) 855-0440
>
>
>
Received on Mon Jul 23 2007 - 20:08:29 CEST