Re: Problem with CopyTree and writing to file

Sorry for this very late reply. Your mail was lost in a long list.

Did you make progress with this problem? If not, would it be possible to get a minimum running system reproducing this problem?

Rene Brun

>
> Hi,
> I have a TChain from which I want to select certain entries, copy them
> to a new TTree, and write the new TTree to a file.
>
> My problem is that when I copy more than a certain number of entries to
> the new tree, the original tree also gets written to the file.
>
> My macro:
>
> #include "Analysis_split.C"
>
> void WriteEtMiss(Long64_t nbrEntries)
> {
> Analysis_split *a = new Analysis_split(0); //made with MakeClass on
> the TChain I //I want to copy
> from.
> TFile *output = new TFile("EtMissCut.root","RECREATE");
> TTree *EtMiss;
> EtMiss = a->fChain->CopyTree("ETmiss>500","",nbrEntries,0);
> EtMiss->SetName("EtMiss");
> EtMiss->SetTitle("EtMissTitle");
> EtMiss->Write();
> }
>
> The output from root when scanning 5336 events or less from the original
> tree:
>
> root [5] WriteEtMiss(5336)
> root [6] .ls
> TFile** EtMissCut.root
> TFile* EtMissCut.root
> OBJ: TTree EtMiss EtMissTitle : 0 at: 0xcc5ced8
> KEY: TTree EtMiss;1 EtMissTitle
> root [7] EtMiss->GetEntries()
> (long long)2291
> root [8]
>
> This works as I meant it to. The tree EtMiss gets written to the file.
> However, if I try to scan more than 5336 entries, the tree/chain that I
> copy from (t3333) also gets written to the file:
>
> root [1] WriteEtMiss(5337)
> root [2] .ls
> TFile** EtMissCut.root
> TFile* EtMissCut.root
> OBJ: TTree EtMiss EtMissTitle : 0 at: 0xbb4e260
> KEY: TTree t3333;1 Combined Ntuple
> KEY: TTree EtMiss;1 EtMissTitle
> root [3] t3333->GetEntries()
> (long long)2292
> root [4] EtMiss->GetEntries()
> (long long)2292
>
> The chain that I run on has 25256 events. Each file in the chain has 400
> entries (except for the last). If I run on more events, for instance
> 20000, t3333;1 is replaced by t3333;4.
>
> If I run on another dataset, the number for entries where this happens
> is different.
> The phenomenon then occurs at 23384 entries. But the number of entries
> copied to EtMiss is now 2107, which is of the same order as in the first
> example (the second data sample has fewer entries that satisfies the
> selection criteria).
>
> I have never used CopyTree before, so I might be doing something the
> wrong way here.
>
> I use root version 4.02/00 on SLC3. I downloaded binaries from the root
> homepage.
>
> Thanks in advance!
>
> cheers,
> Ola
>
>
>
Received on Fri Feb 11 2005 - 22:17:41 MET