Profile histogram for more than 3 quantities

Hey everybody,

I tried it for a long time now, but I just get stuck each time I think that’s it. Ok, my problem in short terms is just to produce something simillar to a 2D profile but for 3D, which means I want the mean value of a given quantity (say E) in my data tree in dependence of THREE other quantities (say a,b,c), also out of my tree.

One way I tried to produce such an Object was to build several 2D profiles for for various given values of one of the other THREE quantities (say for fixed c) and then to replace the BinContents of a given TH3 (that has been built out of a,b,c) by the values of coming out of each of the 2D profiles with respect to c.

This should give me a 3 dimensional histogram where I can “read off” the desired just by means of the BinContent in dependece of a,b,c. But for various reasons this isn’t going to work.

Anyway, can someone of you imagine how to produce an object similar to s.th. that s.o. could call a “3D profile”? If yes, please let me know.

Thanks a lot,
best,

Chris

Well! your description is pretty clear. You want to see a TProfile3D class :smiley:

Rene

I tried to find a TProfile3D class in your reference guide but there is no such a thing. I thought Root is just able to produce up to TProfile2D classes? If you could tell me where to find such a class I would highly appreciate it.

thanks,
chris

Or is there maybe a way to define such a class by my own?

Is it possible for you guys to write a new class for a TProfile3D? This would make things much more easier…

thanks,
Chris

see class TProfile3D now in CVS

root.cern.ch/root/htmldoc//TProfile3D.html

Rene

Absolute fantastic :smiley:
I try it at once…

thanks a lot,
Chris

ok, this might sound somehow stupid, but how can I build this TProfile3D class into my existing ROOT version, or how to update to newest CVS?

Chris

Have a look at

root.cern.ch/root/CVS.html

Cheers, Mathias

hey Rene,

finally I got the TProfile3D to be running on my machine. It works pretty good so far.

thank you very much for that :smiley: ,
best,
Chris