I will take it from the beginning:
I produce a file (file.root) with a number of branches which I then add into a TChain. The tree has e number of entries. This is no problem.
But now I make 10 files that are identical in structure (files*.root) and I would like to merge them so that I still in the end only have one tree. I.e I don’t wish to have 10 times the number of entries which I get from adding 10 files to the chain. I have been trying to do this using TChain::Merge(). I manage to produce a file “all.root” with this code:
//Chain
m_chainName = “voxels”;
m_chain = new TChain(m_chainName);
m_chain->Add(files*.root);
//Input TTree
AnalysisTree = new FluenceNtuple(m_chain); //Load FluenceNtuple-Object with tree "voxels"
cout<<“File: “<<filename1<<” loaded”<<endl;
m_chain->Merge(“all.root”);
But with a few problems:
When I check the number of entries it is 10 times the initial number
all.root gets very big (larger than the sum of the initial files)
How can I achieve what I wish?
Best Regards
Karin
using root 4.03/02 on windows xp
I cannot reproduce your problem. I tried the two ways
root [0] TChain ch("voxels")
root [1] ch.Add("Pri*.root")
(Int_t)3
root [2] ch.Merge("all2.root")
In both cases, the merged Tree has 167343 entries and is 19120175 bytes
Could you tell me exactly what you are doing?
I have a Macro called TFluenceTool.C. One of the functions is:
void TFluenceTool::MergeFile(TString filename)
{
//Chain
m_chainName = “voxels”;
m_chain = new TChain(m_chainName);
m_chain->Add(filename);
//Input TTree
AnalysisTree = new FluenceNtuple(m_chain);
m_chain->Merge(“all.root”);
}
I then run the script by writing .x run.C in the root window, fluence.C is:
void run()
{
gROOT->LoadMacro( “TFluenceTool.C” );
TFluenceTool* test = new TFluenceTool();
test->MergeFile(“Rene/Prim*.root”);
}
but it does not work at all… The files Rene/Prim*.root are loaded and the file all.root is created, but empty. Is it at least a bit more clear what I am doing? Maybe also where it goes wrong?
Have I understood hadd wrong? I thought it is to add histograms only. What I wish is an NTuple Tree that I can pick nodes from. I can therefore not remove the line “AnalysisTree = new FluenceNtuple(m_chain);”, or is there an other way to do the same thing without the Tree?
I have now tried the hadd and as far as I understand it it does not do what I want to do. I think maybe I have not been clear enough when I explain what I want to do. I should not have used the word “merge” at all. I will try again.
When I use hadd, I do indeed get a file that is a merge of two input files, but just put one after the other. I.e, when I do GetEntrise I get 2*the number of entries in each file. I have now learned that this is what “merge” means, so it is not so strange that this is what happends.
I wish to add the content of the leaves so that I in hte new tree have the same number of entries which are sums of the input files. The values will hence be about 2 times as large for 2 input files (if they have similar values stored), but the number of entries is the same.
It could maybe work, but seems to be a very complicated way, and would it really be faster? The main problem för me now is that it takes very long time to go through all the events when I add threes.
As I see it I will have to construct a large tree with brute force, looping throiugh all my data files and adding values into nodes.
/Karin
[quote]I wish to add the content of the leaves so that I in hte new tree have the same number of entries which are sums of the input files. The values will hence be about 2 times as large for 2 input files (if they have similar values stored), but the number of entries is the same.
[/quote]With do not provide any tool to do this automatic (the major hurdle would probably to come up with a generic meaning of ‘add’ … adding really depends on the ‘semantic’ of the content of the tree).
[quote]As I see it I will have to construct a large tree with brute force, looping throiugh all my data files and adding values into nodes. [/quote]It is still confusing to me that you use the word ‘large’. If I understand what you want the resulting tree should
have the same number of leaf
have the same number of entries
and hence is the same size as any of the input files!.
And yes your best choice is to:[code]
open one of the file, read the tree (maintree), and clone it (newtree).
open all the other files and read the other tree (tree2-10).
loop over each entry
do maintree->GetEntry(i)
add to the value in maintree the values in tree2-10)
The files Rene/Prim*.root are loaded and the file all.root is created, but empty. Is it at least a bit more clear what I am doing? Maybe also where it goes wrong?