Hi Salvatore,
The code you sent me seems to prevent all the flaws I have been mentioning.
Your TopNtuplePlots::SetTheAddress is NOT calling SetBranchAddress and in
consequence will
NOT properly work on a chain. (The branch object belong to the underlying
tree and do not communicate any information to the next tree in the chain)
++i; if ( thT->GetBranch("evt") )
thT->GetBranch("evt")->SetAddress(&evtA);
Should simply be
++i; if ( thT->GetBranch("evt") ) thT->SetBranchAddress("evt",&evtA);
And SetTheAddress should be called on the chain NOT a a tree:
SetTheAddresses(chain);
Cheers,
Philippe.
-----Original Message-----
From: Salvatore Rappoccio [mailto:rappoccio@heplcdf2.harvard.edu]
Sent: Friday, November 19, 2004 1:49 PM
To: Philippe Canal
Cc: rappocc@fnal.gov; roottalk@pcroot.cern.ch
Subject: RE: [ROOT] Can't strip TChain
Hi, Philippe,
I have attached my code to create my base class. The files are:
HighLevelObjects : Class definitions of our tree classes
TopNtuplePlots : Driver base class
TMYtools : Miscellaneous tools
Thanks for taking a look at this. I'll try to update to 4.01/02 if I can.
Thanks,
Sal
On Fri, 19 Nov 2004, Philippe Canal wrote:
> > Okay, I removed it and it still loops over all the files fine, but
> > reading
> back in the ntuple, it only has events from the first file, and the
> other entries are garbage.
>
> Did you really make sure that the addresses are set on the chain via a
> call to chain->SetBranchAddresses?
> If it is the case, either update to ROOT 4.01/02 or later or send me a
> complete example.
>
> Cheers,
> Philippe.
>
> -----Original Message-----
> From: Salvatore Rappoccio [mailto:rappoccio@heplcdf2.harvard.edu]
> Sent: Friday, November 19, 2004 1:04 PM
> To: Philippe Canal
> Cc: rappocc@fnal.gov; roottalk@pcroot.cern.ch
> Subject: RE: [ROOT] Can't strip TChain
>
> Okay, I removed it and it still loops over all the files fine, but
> reading back in the ntuple, it only has events from the first file,
> and the other entries are garbage.
>
> Any ideas how to fix this?
>
> Thanks for all your help, I appreciate it!
>
> Sal
>
>
>
> On Fri, 19 Nov 2004, Philippe Canal wrote:
>
> > Hi,
> >
> > Obviously the one inside the loop Clone the tree over and over
> > again, losing the previous information!
> > Remove it!
> >
> > Cheers,
> > Philippe.
> >
> > -----Original Message-----
> > From: Salvatore Rappoccio [mailto:rappoccio@heplcdf2.harvard.edu]
> > Sent: Friday, November 19, 2004 12:45 PM
> > To: Philippe Canal
> > Cc: rappocc@fnal.gov; roottalk@pcroot.cern.ch
> > Subject: RE: [ROOT] Can't strip TChain
> >
> > Hi, Philippe,
> >
> > I put them in both places:
> >
> >
> >
> > SetTheAddresses(chain); // set the address of ALL the top
> > branches on the chain an NOT on the tree
> > if ( stripTree ) {
> > std::cout << "Setting up strip tree" << std::endl;
> > histf->cd();
> > tskim = (TTree*)chain->GetTree()->CloneTree(0);
> > }
> > prevTreeNum=-1;
> > //Loop over the events
> > for(int ev=firstEvent; ev<nEvents; ev++) {
> > counter++;
> >
> >
> > if ( chain->GetTreeNumber() != prevTreeNum ) {
> > SetTheAddresses(chain->GetTree());
> > prevTreeNum = chain->GetTreeNumber();
> > if ( stripTree ) {
> > histf->cd();
> > tskim->AutoSave();
> > tskim = (TTree *)chain->GetTree()->CloneTree(0);
> > }
> >
> > }
> >
> >
> > ....
> > }
> >
> > Thanks,
> > Sal
> >
> > > Cheers,
> > > Philippe.
> > >
> > > -----Original Message-----
> > > From: Salvatore Rappoccio [mailto:rappoccio@heplcdf2.harvard.edu]
> > > Sent: Friday, November 19, 2004 12:13 PM
> > > To: Philippe Canal
> > > Cc: rappocc@fnal.gov; roottalk@pcroot.cern.ch
> > > Subject: RE: [ROOT] Can't strip TChain
> > >
> > > Hi, Philippe,
> > >
> > > I tried this, and now it seems that I only get the events from the
> > > LAST file of the original chain.
> > >
> > > I also tried adding
> > >
> > > SetTheAddresses(chain); // set the address of ALL the top
> > > branches on the chain an NOT on the tree
> > > if ( stripTree ) {
> > > std::cout << "Setting up strip tree" << std::endl;
> > > histf->cd();
> > > tskim->AutoSave();
> > > tskim = (TTree*)chain->GetTree()->CloneTree(0);
> > > }
> > >
> > >
> > > How do I tell the chain written to histf to pick up the earlier data?
> > >
> > > Thanks,
> > > Sal
> > >
> > > >
> > > > Cheers,
> > > > Philippe.
> > > >
> > > > -----Original Message-----
> > > > From: Salvatore Rappoccio
> > > > [mailto:rappoccio@heplcdf2.harvard.edu]
> > > > Sent: Friday, November 19, 2004 10:31 AM
> > > > To: Philippe Canal
> > > > Cc: roottalk@pcroot.cern.ch
> > > > Subject: RE: [ROOT] Can't strip TChain
> > > >
> > > > Hi, Philippe,
> > > >
> > > > I'm using 3.05/07. I'll give this a try.
> > > >
> > > > Thanks!
> > > > Sal
> > > >
> > > > On Thu, 18 Nov 2004, Philippe Canal wrote:
> > > >
> > > > > Hi Salvatore,
> > > > >
> > > > > Which version of ROOT are using?
> > > > >
> > > > > The problem is that you seemingly are setting the addresses of
> > > > > the underlying original tree but not the addresses of the
> > > > > cloned
> tree:
> > > > >
> > > > > if ( chain->GetTreeNumber() != prevTreeNum ) {
> > > > > SetTheAddresses(chain->GetTree());
> > > > > prevTreeNum = chain->GetTreeNumber();
> > > > > }
> > > > >
> > > > > In recent version of ROOT, it is sufficient to set the address
> > > > > of the original chain (using SetBranchAddress).
> > > > >
> > > > > Cheers,
> > > > > Philippe.
> > > > >
> > > > > -----Original Message-----
> > > > > From: owner-roottalk@pcroot.cern.ch
> > > > > [mailto:owner-roottalk@pcroot.cern.ch]
> > > > > On Behalf Of Salvatore Rappoccio
> > > > > Sent: Thursday, November 18, 2004 2:34 PM
> > > > > To: roottalk@pcroot.cern.ch
> > > > > Subject: [ROOT] Can't strip TChain
> > > > >
> > > > > Hi,
> > > > >
> > > > > I am trying to read a subset of the events in my TChain object
> > > > > to another file, given some selection criteria.
> > > > >
> > > > > I can write the objects out without complaint, but when I try
> > > > > to read back the events that I wrote from the second file of
> > > > > the TChain, I get nonsense values.
> > > > >
> > > > > Any ideas?
> > > > >
> > > > > My code is:
> > > > >
> > > > > {
> > > > > std::cout << "Getting trees" << std::endl;
> > > > >
> > > > > TChain * chain = GetChain();
> > > > > TTree * tskim = 0;
> > > > > int nskim = 0;
> > > > > int counter = 0;
> > > > > int prevTreeNum = -1;
> > > > > int prevRunNum = -1;
> > > > > // Load the tree and set the addresses for the second time
> > > > > chain->LoadTree(0);
> > > > > chain->GetEntry(0);
> > > > >
> > > > > if ( stripTree ) {
> > > > > std::cout << "Setting up strip tree" << std::endl;
> > > > > histf->cd();
> > > > > tskim = (TTree*)chain->GetTree()->CloneTree(0);
> > > > > }
> > > > >
> > > > >
> > > > > prevTreeNum=-1;
> > > > >
> > > > > //Loop over the events
> > > > > for(int ev=firstEvent; ev<nEvents; ev++) {
> > > > > counter++;
> > > > > int lflag = -1;
> > > > > lflag = chain->LoadTree(ev);
> > > > >
> > > > > // If this tree hasn't already been set, set the branch
> addresses
> > > > > if ( chain->GetTreeNumber() != prevTreeNum ) {
> > > > > SetTheAddresses(chain->GetTree());
> > > > > prevTreeNum = chain->GetTreeNumber();
> > > > > }
> > > > >
> > > > > if ( stripTree )
> > > > > lflag = chain->GetEntry(ev,1);
> > > > > else
> > > > > lflag = chain->GetEntry(ev);
> > > > >
> > > > > if ( lflag < 0 ) {
> > > > > cout << "Error in loading trees" << endl;
> > > > > return;
> > > > > }
> > > > >
> > > > > // Decide to strip the event or not
> > > > > bool stripEvent = false;
> > > > >
> > > > >
> > > > > // If this event passes simple cuts, process the
> > > > > // event, returning strip status
> > > > > if ( PassCuts( ev, lflag ) ) {
> > > > > stripEvent = Process( ev, lflag );
> > > > > // If desired, strip this event and write to file
> > > > > if ( stripTree ) {
> > > > > tskim->Fill();
> > > > > nskim++;
> > > > > if ( nskim % 10000 == 0 ) {
> > > > > cout << "Writing to strip tree : nskim = " << nskim << endl;
> > > > > tskim->AutoSave();
> > > > > }
> > > > > }
> > > > > }
> > > > > }
> > > > >
> > > > > // Write stuff to output file
> > > > > histf->cd();
> > > > >
> > > > > // Strip tree
> > > > > if ( stripTree ) {
> > > > > cout << "Writing to strip tree : nskim = " << nskim << endl;
> > > > > tskim->AutoSave();
> > > > > }
> > > > > histf -> Close();
> > > > >
> > > > > }
> > > > >
> > > > >
> > > > > Thanks,
> > > > > Sal Rappoccio
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
>
>
>
This archive was generated by hypermail 2b29 : Sun Jan 02 2005 - 05:50:10 MET