How to copy one histogram from the subdirectory to my canvas

michael91 · April 18, 2016, 9:33am

Hey all,

I have one question I could not find the answer to.
I have a root file with 8 subdirectories. Each subdirectory contains a histogram named “h_iEta_hitADC”.

I want to copy this histogram from each subdirectory and put it on one canvas.

Could anyone please help me?

Here code:


using namespace std;

int canvases_working()

{
  TFile *f = new TFile("analyzed.root");
  //TFile *f2 = new TFile("spuriousSignalCapture2_dataTree.root");
  
  
  
  // Creating canvases
  
  
  hitADC=new TCanvas("canvas1","Run1 hitADC results 13/04/2016", 1024, 960);
  hitPos=new TCanvas("canvas2","Run2 hitPos results 13/04/2016", 1024, 960);
 
  hitADC->Divide(2,4);
  hitPos->Divide(2,4);

  
  //Creating histograms

  char histoname [20];
  char histotitle [20];
  char dirname [20];


  TH1D* hadc[8];
  TH1D* hpos[8];
  
  //loop over the subdirs

  for (Int_t i=1; i<9; i++)
  {
    sprintf(histoname,"histo no:%d",i);
    sprintf(histotitle,"eta no:%d",i);
    sprintf(dirname,"SectorEta%d",i);
   
     
    f->cd(dirname);
    cout<<"dir changed to: "<<dirname<<endl;
   

    hitADC->cd(i);
    hitPos->cd(i);

    //hit ADCs
    hadc[i] = f->Get("h_iEta_hitADC");
   
    cout<<"taken"<<endl;
    hadc[i]->Draw("");
    
    //hit Pos
   /* hpos[i]= new TH1D(histoname,histotitle,1000,0,1000);
    hpos[i]->Draw("");*/


  }
  
}

Thanks in advance!

couet · April 18, 2016, 11:34am

a few fixes:

using namespace std;

int canvases_working()

{
  TFile *f = new TFile("analyzed.root");
  //TFile *f2 = new TFile("spuriousSignalCapture2_dataTree.root");
  
  
  
  // Creating canvases
  
  
  hitADC=new TCanvas("canvas1","Run1 hitADC results 13/04/2016", 1024, 960);
  hitPos=new TCanvas("canvas2","Run2 hitPos results 13/04/2016", 1024, 960);
 
  hitADC->Divide(2,4);
  hitPos->Divide(2,4);

  
  //Creating histograms

  char histoname [20];
  char histotitle [20];
  char dirname [20];


  TH1D* hadc[8];
  TH1D* hpos[8];
  
  //loop over the subdirs

  for (Int_t i=1; i<9; i++)
  {
    sprintf(histoname,"histo no:%d",i);
    sprintf(histotitle,"eta no:%d",i);
    sprintf(dirname,"SectorEta%d",i);
   
    f->cd(dirname);
    cout<<"dir changed to: "<<dirname<<endl;
   
    //hit ADCs
   hitADC->cd(i);
   hadc[i] = f->Get("h_iEta_hitADC");
   hadc[i]->Draw("");
    
    //hit Pos
   hitPos->cd(i);
   hpos[i]= f-Get(...); ???
   hpos[i]->Draw("");

  }
  
}

michael91 · April 18, 2016, 11:51am

Hey,

thanks but for your reply but it still does not work.


using namespace std;

int canvases_working_fix()

{
  TFile *f = new TFile("analyzed.root");

  
  
  
  // Creating canvases
  
  
  hitADC=new TCanvas("canvas1","Run1 hitADC results 13/04/2016", 1024, 960);
  
 
  hitADC->Divide(2,4);
  
  //Creating histograms

  char histoname [20];
  char histotitle [20];
  char dirname [20];


  TH1D* hadc[8];
  
  
  //loop over the subdirs

  for (Int_t i=1; i<9; i++)
  {
    sprintf(histoname,"histo no:%d",i);
    sprintf(histotitle,"eta no:%d",i);
    sprintf(dirname,"SectorEta%d",i);
   
    f->cd(dirname);
    cout<<"dir changed to: "<<dirname<<endl;
   
    //hit ADCs
   hitADC->cd(i);
   hadc[i] = f->Get("h_iEta_hitADC");
   hadc[i]->Draw("");
    
   }
  }

I attach my .root file

The output is:

Error: illegal pointer to class object hadc[i] 0x0 1714 canvases_working_fix.cpp:45:

couet · April 18, 2016, 12:15pm

it seems retrieving the TH1F from the file gives problem. It could be that you file is corrupted ?

oot [0]    TFile *f = new TFile("analyzed.root");
root [1]    TH1F* hadc;
root [2]    f->cd("SectorEta1");
root [3]    hadc = (TH1F*)f->Get("h_iEta_hitADC");
root [4] hadc
(TH1F *) nullptr
root [5]

Pepe_Le_Pew · April 18, 2016, 12:39pm

Try: [code]#include “TFile.h”
#include “TDirectory.h”
#include “TCanvas.h”
#include “TH1.h”
#include “TString.h”

#include
#include

int canvases_working() {
TFile *f = new TFile(“analyzed.root”);
//TFile *f2 = new TFile(“spuriousSignalCapture2_dataTree.root”);

// Creating canvases

TCanvas *hitADC =
new TCanvas(“canvas1”, “Run1 hitADC results 13/04/2016”, 1024, 960);
hitADC->Divide(2, 4);

TCanvas *hitPos =
new TCanvas(“canvas2”, “Run2 hitPos results 13/04/2016”, 1024, 960);
hitPos->Divide(2, 4);

//Creating histograms

char histoname[20];
char histotitle[20];
char dirname[20];

TH1F* hadc[8];
TH1F* hpos[8];

//loop over the subdirs

for (Int_t i = 1; i < 9; i++)
{
sprintf(histoname, “histo no:%d”, i);
sprintf(histotitle, “eta no:%d”, i);
sprintf(dirname, “SectorEta%d”, i);

  std::cout << "dir changed to: " << dirname << std::endl;
  f->cd(dirname);
  
  //hit ADCs
  hitADC->cd(i);
  gDirectory->GetObject(TString::Format("h_iEta%d_hitADC", i),
                        hadc[(i - 1)]);
  if (hadc[(i - 1)]) hadc[(i - 1)]->Draw("");
  
  //hit Pos
  hitPos->cd(i);
  gDirectory->GetObject(TString::Format("h_iEta%d_hitPos", i),
                        hpos[(i - 1)]);
  if (hpos[(i - 1)]) hpos[(i - 1)]->Draw("");
}

hitADC->cd(0);
hitADC->Modified(); hitADC->Update();

hitPos->cd(0);
hitPos->Modified(); hitPos->Update();

return 0;
}[/code]

michael91 · April 18, 2016, 1:13pm

Thank both of you for your help!

It works fine now.

Btw. Is there any good tutorial I could get through to learn as much as possible?

Best regards.

Pepe_Le_Pew · April 18, 2016, 2:44pm

I “simplified” the code: [code]#include “TFile.h”
#include “TDirectory.h”
#include “TCanvas.h”
#include “TH1.h”
#include “TString.h”

#include

int canvases_working() {
TFile *f = new TFile(“analyzed.root”);
//TFile *f2 = new TFile(“spuriousSignalCapture2_dataTree.root”);

// Creating canvases

TCanvas *hitADC =
new TCanvas(“canvas1”, “Run1 hitADC results 13/04/2016”, 1024, 960);
hitADC->Divide(2, 4);

TCanvas *hitPos =
new TCanvas(“canvas2”, “Run2 hitPos results 13/04/2016”, 1024, 960);
hitPos->Divide(2, 4);

//Creating histograms

TH1F* hadc[8];
TH1F* hpos[8];

//loop over the subdirs

for (Int_t i = 1; i < 9; i++)
{
std::cout << "subdir " << i << std::endl;
f->cd(TString::Format(“SectorEta%d”, i));

  //hit ADCs
  hitADC->cd(i);
  gDirectory->GetObject(TString::Format("h_iEta%d_hitADC", i),
                        hadc[(i - 1)]);
  if (hadc[(i - 1)]) hadc[(i - 1)]->Draw("");
  
  //hit Pos
  hitPos->cd(i);
  gDirectory->GetObject(TString::Format("h_iEta%d_hitPos", i),
                        hpos[(i - 1)]);
  if (hpos[(i - 1)]) hpos[(i - 1)]->Draw("");
}

hitADC->cd(0);
hitADC->Modified(); hitADC->Update();

hitPos->cd(0);
hitPos->Modified(); hitPos->Update();

return 0;
}[/code]

michael91 · April 19, 2016, 11:57am

Hey,

In my case everything works fine. If you have any solution of your problems, it would be nice if you could share it.

Thanks for your help!

pcanal · April 20, 2016, 2:57pm

Hi Pepe,

Do you have a copy of ‘analyzed.root’ that you use to reproduce the problem?

Thanks,
Philippe.

Pepe_Le_Pew · April 20, 2016, 3:31pm

.

pcanal · April 20, 2016, 5:05pm

Hi Pepe,

I think the problem is that the code is faulty in regard to index range:TH1F* hadc[8]; .... for (Int_t i = 1; i < 9; i++) ... set hadc[i]

thus hadc[8] is set by the ‘last’ iteration but this is out of bound for hadc which has only 8 elements (0 throught 7) and actually set the value of i. This is not seen in the std::cout output ‘thanks’ to optimization to use a register for the printing but use a memory address for the comparison.

Cheers,
Philippe.

Pepe_Le_Pew · April 20, 2016, 5:13pm

Many thanks for finding this bug (I wonder why neither the interpreter nor the compiler complained about it and there was no “out of range” / “index out of bounds” error when running it -> I guess I’m getting too old that I did not spot it myself).

I corrected the source code in both my previous posts.

Pepe_Le_Pew · April 20, 2016, 6:27pm

Philippe,
for future reference, could you, please, say how you “detected” this bug.

I tried:
valgrind --tool=memcheck --leak-check=full --suppressions=root-config --etcdir/valgrind-root.supp root-config --bindir/root.exe -l -q canvases_working.cxx++

For ROOT 5 (v5-34-00-patches “debug”), it unfortunately exited without problems and I got one lost block related to std::string / G__setup_func_struct and another six lost blocks related to libfontconfig / libXft / TGX11TTF.

For ROOT 6 (6.06/02 “RelWithDebInfo”), the “for” loop did not stop at 8, so I had to kill it with CTRL+C, and then I got hundreds of lost blocks, so it’s not really possible to analyse it (well, I checked several in beginning and in the end).

pcanal · April 20, 2016, 6:51pm

Hi Pepe,

[quote]For ROOT 6 I got hundreds of lost blocks, so it’s not really possible to analyse it (well, I checked several in beginning and in the end).[/quote]Turning on leak check would not help whatsoever here, as it just detect memory leak and not memory override. As a side note, I am in the tail end of a major set of updates to the suppression files that should illuminate most (hopefully) all the (per-se) false-positive report of possibly-lost.

Unfortunately, valgrind is not properly detecting memory overwrite when the memory is on the stack and indeed I ran valgrind and it did not report any issues.

I discovered the problem by code inspection of the code I ended-up when trying to reducing the problem to be minimal set. I end up with:[code]#include

class Histo { public: void Print() { } };

int canvases_working() {
Histo* hadc[9];

//loop over the subdirs
int counter = 0;
for (Int_t iii = 1; iii < 9; iii++)
{
std::cout << "subdir " << iii << std::endl;

  hadc[iii] = nullptr;
  if (hadc[iii]) hadc[iii]->Print();
 
  if (counter > 10) return 2;
  ++counter;
}

return 0;
}[/code].

I also had a look at the assembly code produced (via gdb) to see what’s was going on (hence my affirmation on register vs memory).

Cheers,
Philippe.

Pepe_Le_Pew · April 20, 2016, 7:03pm

O.K. Thanks (so, you did the same as I usually do -> just in this case I forgot about it completely).

I also tried (“exp-sgcheck” is supposed to deal with “stack and global array overrun” problems):
valgrind --tool=exp-sgcheck --suppressions=root-config --etcdir/valgrind-root.supp root-config --bindir/root.exe -l -q canvases_working.cxx++

It seems to me that “exp-sgcheck” completely refuses to work with ROOT (I tried v5-34-00-patches “debug” and 6.06/02 “RelWithDebInfo”) -> it “immediately” exits without running the macro: [code][…]$ valgrind --tool=exp-sgcheck --suppressions=root-config --etcdir/valgrind-root.supp root-config --bindir/root.exe -l -q canvases_working.cxx++
==966== exp-sgcheck, a stack and global array overrun detector
==966== NOTE: This is an Experimental-Class Valgrind Tool
==966== Copyright © 2003-2015, and GNU GPL’d, by OpenWorks Ltd et al.
==966== Using Valgrind-3.12.0.SVN and LibVEX; rerun with -h for copyright info
==966== Command: /opt/ROOT/debug/v5-34-00-patches/bin/root.exe -l -q canvases_working.cxx++
==966==
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93
–966-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0x93

exp-sgcheck: sg_main.c:559 (add_blocks_to_StackTree): Assertion ‘!already_present’ failed.

host stacktrace:
==966== at 0x38013B88: show_sched_status_wrk (m_libcassert.c:343)
==966== by 0x38013C94: report_and_quit (m_libcassert.c:415)
==966== by 0x38013E21: vgPlain_assert_fail (m_libcassert.c:481)
==966== by 0x3800A980: add_blocks_to_StackTree (sg_main.c:559)
==966== by 0x3800B8A9: shadowStack_new_frame.isra.22 (sg_main.c:1883)
==966== by 0x805092FE9: ???
==966== by 0x804072F2F: ???
==966== by 0x80206456F: ???
==966== by 0x61E757A: __nscd_get_mapping (nscd_helper.c:286)
==966== by 0x80206456F: ???
==966== by 0x1BFF: ???

sched status:
running_tid=1

Thread 1: status = VgTs_Runnable (lwpid 966)
==966== at 0x61E75B8: __nscd_get_mapping (nscd_helper.c:293)
==966== by 0x61E7ABB: __nscd_get_map_ref (nscd_helper.c:443)
==966== by 0x61E4207: nscd_getpw_r (nscd_getpw_r.c:95)
==966== by 0x61E4645: __nscd_getpwuid_r (nscd_getpw_r.c:63)
==966== by 0x617538F: getpwuid_r@@GLIBC_2.2.5 (getXXbyYY_r.c:196)
==966== by 0x6174B85: getpwuid (getXXbyYY.c:116)
==966== by 0x517A11C: TUnixSystem::UnixHomedirectory(char const*) (TUnixSystem.cxx:4032)
==966== by 0x5173FCF: TUnixSystem::HomeDirectory(char const*) (TUnixSystem.cxx:1509)
==966== by 0x50BDD8C: TROOT::InitSystem() (TROOT.cxx:1434)
==966== by 0x50B8960: TROOT::TROOT(char const*, char const*, void (**)()) (TROOT.cxx:298)
==966== by 0x50B7EEE: ROOT::GetROOT() (TROOT.cxx:206)
==966== by 0x50C0334: __static_initialization_and_destruction_0(int, int) (TROOT.cxx:215)
==966== by 0x50C037A: _GLOBAL__sub_I_TROOT.cxx (TROOT.cxx:2236)
==966== by 0x4010139: call_init.part.0 (dl-init.c:78)
==966== by 0x4010222: _dl_init (dl-init.c:36)
==966== by 0x4001309: ??? (in /lib/x86_64-linux-gnu/ld-2.19.so)
==966== by 0x3: ???
==966== by 0xFFEFFFE1A: ???
==966== by 0xFFEFFFE48: ???
==966== by 0xFFEFFFE4B: ???
==966== by 0xFFEFFFE4E: ???

Note: see also the FAQ in the source distribution.
It contains workarounds to several common problems.
In particular, if Valgrind aborted or crashed after
identifying problems in your program, there’s a good chance
that fixing those problems will prevent Valgrind aborting or
crashing, especially if it happened in m_mallocfree.c.

If that doesn’t help, please report this bug to: www.valgrind.org

In the bug report, send all the above text, the valgrind
version, and what OS and version you are using. Thanks.[/code] Actually, it looks like it is sufficient to #include “TObject.h” (or anything that inherits from TObject) and the “exp-sgcheck” is dead (you do not need to create any object, just #include the header file).
On the other hand, if I just #include “TString.h” then “exp-sgcheck” works fine (does not need any ROOT libraries). However, as soon as I add “TString s;” in the source code, then it dies (needs libCore and libCint ROOT 5 libraries or just libCore in ROOT 6).
So, the problem is somehow related to ROOT libraries.

I tried Valgrind-3.10.1 (the default version on Ubuntu 14.04.4 LTS), Valgrind 3.11.0 (the current release) and the current svn trunk, and in all cases I get the same problem.

pcanal · April 21, 2016, 5:50pm

Hi Pepe,

This ‘sounds’ like it is a problem in sgcheck. Can you reduce the problem even more (i.e. which part of TObject.h is triggering this)?

Thanks,
Philippe

Pepe_Le_Pew · April 22, 2016, 6:16am

In the end … it turns out that, in order to collapse “exp-sgcheck”, one does not even need to #include anything from ROOT … see a new thread dedicated to this problem: viewtopic.php?f=7&t=21520

P.S. See also another thread which may be helpful in finding use-after-free and {heap,stack,global}-buffer overflow bugs: viewtopic.php?f=7&t=21522