void runmergeWithDir()
#endif
{
TChain *tph;
tph = new TChain("CBNT/t3333");
tph->AddFile("file1.root");
tph->AddFile("file2.root");
tph->Merge("merged.root");
TFile *fil = new TFile("merged.root");
fil->ls();
}
void formChain300(){
cout << "Begin loading..." << endl;
const TString treedirectory = "demo/T_weights";
TChain* InputChain = new TChain(treedirectory);
InputChain->Add("rfio:/castor/cern.ch/user/k/kreis/CUSusy/QCD/T_weight_QCD_Pt300_*.root");
InputChain->Merge("/cu1/kreis/QCD/T_weights_QCD_Pt300.root");
cout << "Loaded " << InputChain->GetListOfFiles()->GetEntries() << " files" << endl;
return;
}
void macro( TString dirname = "thin" )
{
//cout<<"Processing directory:"<<dirname<<endl;
gROOT->Reset();
gROOT->SetStyle("Plain");
gStyle->SetFillColor(0); //white fill color
gStyle->SetFrameBorderMode(0); //no frame border
gStyle->SetCanvasBorderMode(0); //no canvas border
gStyle->SetOptTitle(0); //no title
//gStyle->SetOptStat(0); //no statistics _or_
gStyle->SetOptStat(1111111111);
// -->all statistics (the more 1's y
// connect file and tree
//TFile *file = new TFile(dirname+"IPbinpT_0.4.root");
//TFile *f = (TFile*)gROOT->GetListOfFiles()->FindObject(dirname+"IPbinpT_0.33.root");
//if (f==0) {
// std::cout << "What This file doesn't exist MAAAAN!" << std::endl;
//}
//TTree *tree = (TTree*)gDirectory->Get("sigma");
// Create the chain as above
//
TChain *chain = new TChain("sigma");
chain->Add(dirname+"/*.root");
chain->Merge(dirname+".root");
//cout<<"Merged file: "<<dirname<<".root"<<endl;
//exit();
//chain->Draw("bx1:pTi");
/* for (Int_t i(0); i<9; ++i)
{
chain->GetEvent(i);
}
*/
/*chain.SetBranchAddress("event", &event);
// Start main loop on all events
// In case you want to read only a few branches, use TChain::SetBranchStatus
// to activate/deactivate a branch.
Int_t nevent = chain.GetEntries();
for (Int_t i=0;i<nevent;i++) {
chain.GetEvent(i); //read complete accepted event in memory
hnseg->Fill(event->GetNseg()); //Fill histogram with number of segments
}
// Draw the histogram
hnseg->Draw();*/
}
void WriteMergeObjects( TFile *target ) {
cout << "Writing the merged data." << endl;
TIterator *nextobj = MergeObjects.MakeIterator();
TObjString *pathname_obj;
while( (pathname_obj = (TObjString *)nextobj->Next()) ) {
TString path,name;
SplitPathName(pathname_obj->String(),&path,&name);
TObject *obj = MergeObjects.GetValue(pathname_obj);
target->cd(path);
obj->Write( name );
delete obj;
}
MergeObjects.Clear();
target->Write();
// Temporarily let multiple root files remain if > 2GB
// Prevent Target_1.root Target_2.root, ... from happening.
// long long max_tree_size = 200000000000LL; // 200 GB
// if(TTree::GetMaxTreeSize() < max_tree_size ) {
// TTree::SetMaxTreeSize(max_tree_size);
// }
nextobj = MergeChains.MakeIterator();
TObjString *pathname_obj;
while( (pathname_obj = (TObjString *)nextobj->Next()) ) {
TString path,name;
SplitPathName(pathname_obj->String(),&path,&name);
TChain *ch = (TChain *)MergeChains.GetValue(pathname_obj);
target->cd(path);
ch->Merge(target,0,"KEEP");
delete ch;
// in case of multiple objects with same pathname, must remove
// this one from the list so we don't get the same (deleted)
// one next time we look up the same name
MergeChains.Remove(pathname_obj);
}
MergeChains.Clear();
InitializedMergeObjects = false;
}
void mergeHadoopFiles() {
gSystem->Load("../libMiniFWLite.so");
TChain *chain = new TChain("t");
chain->SetMaxTreeSize(5000000000LL); //default is 100000000000LL
chain->Add("unmerged/ww2l2nujets_merged_ntuple_10_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_11_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_1_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_2_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_3_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_4_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_5_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_6_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_7_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_8_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_9_smallTree.root");
chain->Add("unmerged/ww2l2nujets_merged_ntuple_smallTree.root");
chain->Merge("merged/ww2l2nujets.root", "fast");
}
void mergeHadoopFiles(const TString& indir, const TString& outpath) {
// Benchmark
TBenchmark *bmark = new TBenchmark();
bmark->Start("benchmark");
//gSystem->Load("../libMiniFWLite.so");
TChain *chain = new TChain("Events");
chain->SetMaxTreeSize(5000000000LL); //default is 100000000000LL
std::cout << "Merging files from dir: " << indir << std::endl
<< "Ouputting to: " << outpath << std::endl;
chain->Add(indir + "/*.root");
chain->Merge(outpath, "fast");
std::cout << "Total events merged: " << chain->GetEntries() << std::endl;
bmark->Stop("benchmark");
cout << "------------------------------" << endl;
cout << "CPU Time: " << Form( "%.01f s", bmark->GetCpuTime("benchmark") ) << endl;
cout << "Real Time: " << Form( "%.01f s", bmark->GetRealTime("benchmark") ) << endl;
cout << endl;
delete bmark;
}
void hmerge_( TDirectory *target, TList *sourcelist, double crossArray[] ) {
//cout << "Target path: " << target->GetPath() << endl;
TString rpath( (char*)strstr( target->GetPath(), ":" ) );
rpath.Remove( 0, 2 );
TFile *first_source = (TFile*)sourcelist->First();
first_source->cd( rpath );
TDirectory *current_sourcedir = gDirectory;
//gain time, do not add the objects in the list in memory
Bool_t status = TH1::AddDirectoryStatus();
TH1::AddDirectory(kFALSE);
// loop over all keys in this directory
TChain *globChain = 0;
TIter nextkey( current_sourcedir->GetListOfKeys() );
TKey *key, *oldkey=0;
while ( (key = (TKey*)nextkey())) {
//keep only the highest cycle number for each key
if (oldkey && !strcmp(oldkey->GetName(),key->GetName())) continue;
// read object from first source file
first_source->cd( rpath );
TObject *obj = key->ReadObj();
if ( obj->IsA()->InheritsFrom( "TH1" ) ) {
// descendant of TH1 -> merge it
//cout << "Merging histogram " << obj->GetName() << endl;
TH1 *h1 = (TH1*)obj;
h1->Sumw2();
// Scale by the cross-section factor
h1->Scale(crossArray[0]);
// loop over all source files and add the content of the
// correspondant histogram to the one pointed to by "h1"
TFile *nextsource = (TFile*)sourcelist->After( first_source );
int q = 1; // This keeps track of which
// cross section factor to use
while ( nextsource ) {
// make sure we are at the correct directory level by cd'ing to rpath
nextsource->cd( rpath );
TKey *key2 = (TKey*)gDirectory->GetListOfKeys()->FindObject(h1->GetName());
if (key2) {
TH1 *h2 = (TH1*)key2->ReadObj();
h2->Sumw2();
// Scale by the cross-section factor before adding.
h2->Scale(crossArray[q]);
h1->Add( h2 );
q++;
delete h2;
}
nextsource = (TFile*)sourcelist->After( nextsource );
}
} else if ( obj->IsA()->InheritsFrom( "TTree" ) ) {
// loop over all source files create a chain of Trees "globChain"
const char* obj_name= obj->GetName();
globChain = new TChain(obj_name);
globChain->Add(first_source->GetName());
TFile *nextsource = (TFile*)sourcelist->After( first_source );
while ( nextsource ) {
globChain->Add(nextsource->GetName());
nextsource = (TFile*)sourcelist->After( nextsource );
}
} else if ( obj->IsA()->InheritsFrom( "TDirectory" ) ) {
cout << "Found subdirectory " << obj->GetName() << endl;
// create a new subdir of same name and title in the target file
target->cd();
TDirectory *newdir = target->mkdir( obj->GetName(), obj->GetTitle() );
// newdir is now the starting point of another round of merging
// newdir still knows its depth within the target file via
// GetPath(), so we can still figure out where we are in the recursion
hmerge_( newdir, sourcelist,crossArray );
} else {
// object is of no type that we know or can handle
cout << "Unknown object type, name: "
<< obj->GetName() << " title: " << obj->GetTitle() << endl;
}
// now write the merged histogram (which is "in" obj) to the target file
// note that this will just store obj in the current directory level,
// which is not persistent until the complete directory itself is stored
// by "target->Write()" below
if ( obj ) {
target->cd();
//!!if the object is a tree, it is stored in globChain...
if(obj->IsA()->InheritsFrom( "TTree" ))
globChain->Merge(target->GetFile(),0,"keep");
else
obj->Write( key->GetName() );
}
} // while ( ( TKey *key = (TKey*)nextkey() ) )
// save modifications to target file
target->SaveSelf(kTRUE);
TH1::AddDirectory(status);
}
void MergeRootfile( TDirectory *target, TList *sourcelist) {
// cout << "Target path: " << target->GetPath() << endl;
TString path( (char*)strstr( target->GetPath(), ":" ) );
path.Remove( 0, 2 );
TFile *first_source = (TFile*)sourcelist->First();
first_source->cd( path );
TDirectory *current_sourcedir = gDirectory;
//gain time, do not add the objects in the list in memory
Bool_t status = TH1::AddDirectoryStatus();
TH1::AddDirectory(kFALSE);
// loop over all keys in this directory
TChain *globChain = 0;
TIter nextkey( current_sourcedir->GetListOfKeys() );
TKey *key, *oldkey=0;
std::vector<double>::iterator wItr;
while ( (key = (TKey*)nextkey())) {
wItr=WeightList->begin();
//keep only the highest cycle number for each key
if (oldkey && !strcmp(oldkey->GetName(),key->GetName())) continue;
// read object from first source file
first_source->cd( path );
TObject *obj = key->ReadObj();
if ( obj->IsA()->InheritsFrom( "TH1" ) ) {
// descendant of TH1 -> merge it
// cout << "Merging histogram " << obj->GetName() << endl;
TH1 *h1 = (TH1*)obj;
h1->Sumw2();
// weight is eff * sigma, because the histos are produced on the ~same initial
// events, and eff is already included (in reconstruction), weight is just sigma
h1->Scale(double(*wItr));
// loop over all source files and add the content of the
// correspondant histogram to the one pointed to by "h1"
TFile *nextsource = (TFile*)sourcelist->After( first_source );
wItr++;
while ( nextsource ) {
// make sure we are at the correct directory level by cd'ing to path
nextsource->cd( path );
TKey *key2 = (TKey*)gDirectory->GetListOfKeys()->FindObject(h1->GetName());
if (key2) {
TH1 *h2 = (TH1*)key2->ReadObj();
h2->Sumw2();
h2->Scale( double(*wItr) );
h1->Add( h2 );
delete h2;
}
nextsource = (TFile*)sourcelist->After( nextsource );
wItr++;
}
}
else if ( obj->IsA()->InheritsFrom( "TTree" ) ) {
// loop over all source files create a chain of Trees "globChain"
const char* obj_name= obj->GetName();
globChain = new TChain(obj_name);
globChain->Add(first_source->GetName());
TFile *nextsource = (TFile*)sourcelist->After( first_source );
// const char* file_name = nextsource->GetName();
// cout << "file name " << file_name << endl;
while ( nextsource ) {
globChain->Add(nextsource->GetName());
nextsource = (TFile*)sourcelist->After( nextsource );
}
} else if ( obj->IsA()->InheritsFrom( "TDirectory" ) ) {
// it's a subdirectory
cout << "Found subdirectory " << obj->GetName() << endl;
// create a new subdir of same name and title in the target file
target->cd();
TDirectory *newdir = target->mkdir( obj->GetName(), obj->GetTitle() );
// newdir is now the starting point of another round of merging
// newdir still knows its depth within the target file via
// GetPath(), so we can still figure out where we are in the recursion
MergeRootfile( newdir, sourcelist );
} else {
// object is of no type that we know or can handle
cout << "Unknown object type, name: "
<< obj->GetName() << " title: " << obj->GetTitle() << endl;
}
// now write the merged histogram (which is "in" obj) to the target file
// note that this will just store obj in the current directory level,
// which is not persistent until the complete directory itself is stored
// by "target->Write()" below
if ( obj ) {
target->cd();
//!!if the object is a tree, it is stored in globChain...
if(obj->IsA()->InheritsFrom( "TTree" ))
globChain->Merge(target->GetFile(),0,"keep");
else
obj->Write( key->GetName() );
}
} // while ( ( TKey *key = (TKey*)nextkey() ) )
// save modifications to target file
target->SaveSelf(kTRUE);
TH1::AddDirectory(status);
}
void MergeRootfile( TDirectory *target, const vector<pair<TFile*, float> >& vFileList) {
// cout << "Target path: " << target->GetPath() << endl;
TString path( (char*)strstr( target->GetPath(), ":" ) );
path.Remove( 0, 2 );
vec_pair_it it = vFileList.begin();
TFile *first_source = (*it).first;
const float first_weight = (*it).second;
first_source->cd( path );
TDirectory *current_sourcedir = gDirectory;
//gain time, do not add the objects in the list in memory
Bool_t status = TH1::AddDirectoryStatus();
TH1::AddDirectory(kFALSE);
// loop over all keys in this directory
TChain *globChain = 0;
TIter nextkey( current_sourcedir->GetListOfKeys() );
TKey *key, *oldkey=0;
while ( (key = (TKey*)nextkey())) {
//keep only the highest cycle number for each key
if (oldkey && !strcmp(oldkey->GetName(),key->GetName())) continue;
// read object from first source file
first_source->cd( path );
TObject *obj = key->ReadObj();
if ( obj->IsA()->InheritsFrom( TH1::Class() ) ) {
// descendant of TH1 -> merge it
// cout << "Merging histogram " << obj->GetName() << endl;
TH1 *h1 = (TH1*)obj;
if (first_weight>0) {
h1->Scale(first_weight);
}
// loop over all source files and add the content of the
// correspondant histogram to the one pointed to by "h1"
for (vec_pair_it nextsrc = vFileList.begin()+1; nextsrc != vFileList.end(); ++nextsrc ) {
// make sure we are at the correct directory level by cd'ing to path
(*nextsrc).first->cd( path );
const float next_weight = (*nextsrc).second;
TKey *key2 = (TKey*)gDirectory->GetListOfKeys()->FindObject(h1->GetName());
if (key2) {
TH1 *h2 = (TH1*)key2->ReadObj();
if (next_weight>0) h2->Scale(next_weight);
h1->Add( h2 );
delete h2;
}
}
} else if ( obj->IsA()->InheritsFrom( TTree::Class() ) ) {
// loop over all source files create a chain of Trees "globChain"
const char* obj_name= obj->GetName();
globChain = new TChain(obj_name);
globChain->Add(first_source->GetName());
for (vec_pair_it nextsrc = vFileList.begin()+1; nextsrc != vFileList.end(); ++nextsrc ) {
globChain->Add(nextsrc->first->GetName());
}
} else if ( obj->IsA()->InheritsFrom( TDirectory::Class() ) ) {
// it's a subdirectory
cout << "Found subdirectory " << obj->GetName() << endl;
// create a new subdir of same name and title in the target file
target->cd();
TDirectory *newdir = target->mkdir( obj->GetName(), obj->GetTitle() );
// newdir is now the starting point of another round of merging
// newdir still knows its depth within the target file via
// GetPath(), so we can still figure out where we are in the recursion
MergeRootfile( newdir, vFileList);
} else {
// object is of no type that we know or can handle
cout << "Unknown object type, name: "
<< obj->GetName() << " title: " << obj->GetTitle() << endl;
}
// now write the merged histogram (which is "in" obj) to the target file
// note that this will just store obj in the current directory level,
// which is not persistent until the complete directory itself is stored
// by "target->Write()" below
if ( obj ) {
target->cd();
//!!if the object is a tree, it is stored in globChain...
if(obj->IsA()->InheritsFrom( TTree::Class() ))
globChain->Merge(target->GetFile(),0,"keep");
else
obj->Write( key->GetName() );
}
} // while ( ( TKey *key = (TKey*)nextkey() ) )
// save modifications to target file
target->SaveSelf(kTRUE);
TH1::AddDirectory(status);
}
