[ROOT] Problem with (Option "E3") filled error bands due to empty bins and SetMinimum()

From: Marc Brueggemann (bruegge@physik.uni-dortmund.de)
Date: Wed Jul 10 2002 - 10:09:57 MEST 

Hi everybody,

I realized a problem with drawing option "E3" to draw a colored error
band. It occurs if I apply this option to a histogram which also
contains empty bins and using the method 'SetMinimum()' simultaneously.
Starting from the last bins the error band is focused to the origin.
A strange fragment occurs.  
   
I have reduced the problem to a little macro which you can try out.
It contains two loops for filling a histogram with errors. One loop
leaves the first bins empty the other (uncomment it) fills
all bins.   



Int_t Test(){
  
  TH1F* h_test = new TH1F("h_test","h_test", 45, 0, 12 );
  
  // for (Int_t i = 1; i <= 45; ++i) { // use this loop and no problem
  // occurs (because no empty bins are present)
  for (Int_t i = 6; i <= 45; ++i) {
    h_test->SetBinContent(i, 4.0);
    Float_t error = (40.1-i)/40.0;
    h_test->SetBinError(i, error);
  }
  
  TCanvas* canvas = new TCanvas("canvas", "canvas", 1);
  h_test->SetMinimum(0.01);
  h_test->SetMaximum(6);
  h_test->SetFillColor(kRed);
  h_test->Draw("E3");
  h_test->Draw("P HIST SAME");
  

  return 0;
}


The same fragment occurs if I don't use 'SetMinimum()' but afterwards
(if I can see the histogram in a canvas) 
try to zoom into a certain region of the y-axis.     


The question: Does this problem also occur if you try this
              macro?

Am I doing something wrong?? How can I keep ROOT from connecting the
error band with the origin ???

I'm using ROOT-version 3.02/02 
(I also tried version 2.24/05 and the same problem occured) 



Best regards

  Marc Brueggemann 

--
Marc Brueggemann        mailto:bruegge@Physik.Uni-Dortmund.DE
Experimentelle Physik Vb            phone: +49 (231) 755 3667
Dortmund University
D-44227 Dortmund

This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:50:59 MET