Hi Chris,
Thanks for reporting this problem with the last bin in case of
TH1::Rebin.
I have now fixed the problem in my development version.
Your fix is not sufficient. I have reorganized a bit the logic.
Rene Brun
Chris Jillings wrote:
>
> Hello,
> I think I have found a problem with Rebin. It seems to show itself in
> the last bin of a histogram.
> I have a 100-bin histo, h, which I rebin to 50 with h->Rebin().
> The 50th bin of the histo contains fewer counts than it should. In fact in
> contains counts from the 99th, but not the 100th bin of the original
> histo.
>
> Looking at the html/source file I think I might have found the problem.
>
> Double_t binContent, binError;
> for (bin = 1;bin<=newbins;bin++) {
> binContent = 0;
> binError = 0;
> for (i=0;i<ngroup;i++) {
> if (oldbin+i >= nbins) break; // PROBLEM HERE ?? should be >nbins
> binContent += oldBins[oldbin+i];
> if (oldErrors) binError += oldErrors[oldbin+i]*oldErrors[oldbin+i];
> }
> hnew->SetBinContent(bin,binContent);
> if (oldErrors) hnew->SetBinError(bin,TMath::Sqrt(binError));
> oldbin += ngroup;
> }
>
> Am I correct?
>
> Thanks,
>
> Chris
>
> ==================================================================
> = =
> = Chris Jillings =
> = Department of Physics phone/voice mail: (613) 533-6000x74805 =
> = Queen's University fax: (613) 533-6813 =
> = Kingston, Ontario email: chris@sno.phy.queensu.ca =
> = Canada, K7L 3N6 web: http://sno.phy.queensu.ca =
> = =
> ==================================================================
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:43:35 MET