Binning of TH2D is changed after filling

ROOT
1

Dear experts,

I created a few histograms. But I found after filling a few of entries, the binning of histograms are changed.

The related code are following

    (hSignal_D0[imass][ipt])["largeDCA"] = new TH2D(Form("hSignal_mass%d_pt%d_largeDCA", imass, ipt),
                "", ana::nEtaBin, ana::etaBegin, ana::etaEnd,
                //ana::nPhiBin, ana::phiBegin, ana::phiEnd);
                ana::nPhiBin, -1.57, ana::phiEnd);

     (hSignal_D0[imass][ipt])["smallDCA"] = new TH2D(Form("hSignal_mass%d_pt%d_smallDCA", imass, ipt),
                "", ana::nEtaBin, ana::etaBegin, ana::etaEnd,
                //ana::nPhiBin, ana::phiBegin, ana::phiEnd);
                ana::nPhiBin, -1.57, ana::phiEnd);

 (hSignal_D0[imass][ipt])[strDCA]->Fill(deltaEta, deltaPhi,
                         1./nMult_trg_eff_d0[imass][ipt]/effks/effweight_ass);

    for(int imass=0; imass<ana::nMass; imass++){
       for(int ipt=0; ipt<ana::nPt; ipt++){
          for(auto& h : hSignal_D0[imass][ipt]) {
             std::cout << h.second->GetYaxis()->GetXmin() << std::endl;;
          }
       }
    }

Also I update the log. If you need the source code, I can upload all. But I am not sure whether a few hundred lines of codes are too long.

Many thanks in advance.

log.txt (2.3 KB)

ROOT Version: 6.06/01
Platform: linuxx8664gcc
Compiler: gcc version 5.3.0 (GCC)

2

Are you sure that always “ana::phiEnd > -1.57” (if not, the y-axis limits will be automatically computed).

3

No, I cannot. but would not the smaller value be stored in underflow?
Thanks

4

See “Histograms with automatic bins” (and then search also for “automatic binning”) in the TH1 class reference.

5

Hi, Wile, in my case, I used the GetYaxis()->GetXmax() to show the max value of the axis. Actually, every phiEnd larger than -1.57 and positive. But the max value also be changed after filling a few times. Also I used h.second->SetCanExtend(TH1::kNoAxis), this does not work yet.

6

Can you maybe try the newest ROOT version (maybe it’s a bug in your very old version).

7

The latest version I can use is 6.12/07… I am using cmssw on the cluster… The bug is very annoying. Do you need my macros of full version?
Thanks

8

let’s ping … @couet

9

Yes a running macro reproducing the problem would be very helpful. Just a subset of your program will be sufficient as soon as it shows the issue.

10

Hi, Couet

I just remove all irrelevant parts of my codes. I have attached the source code.

After I commented the hist->Fill(), bug did not happen. Also I attach the link to the datasets
https://drive.google.com/drive/folders/1Pj8TqCs2e47f9cB74mV8LL7qiDHfdxwP?usp=sharing
d0ana_1.root would not give bad binning but d0ana_10.root would.

Many thanks

corr2D.C (6.5 KB)
Event.cxx (3.4 KB)
Event.h (4.3 KB)
myAnaConsts.h (1.4 KB)

11

Hi, Wile,
Sorry for my mistake, I just mistakenly set wrong range of x axis, and y axis changed with x. Thank you!

12

Hi, couet
Sorry for my mistake, I just mistakenly set wrong range of x axis, and y axis changed with x. I only checked y axis range before. Thank you!

13

Yes, in the “myAnaConsts.h” file, you set etaBegin = etaEnd = -4.95 and this triggers the “automatic binning” for the x-axis.

14

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.