Re: Defining TF2 functions and doing fits in for loop

Hi,

All your fit functions have the same name, replacing existing ones. If you use is instead of isize in the call to sprintf it should work.

Cheers, Axel

nguler_at_jlab.org wrote:

>Hello All,
>
>I am using root 5.26-00-64bit. I have a TH2F object:
>
> TH2F *histsim = new TH2F(histsimName,histsimName,256,0,256,256,0,256);
>
>Then I define fit functions:
>
> TF2 *fitsim1_5 =
> new
>TF2("fitsim1_5",Gaus2D,jVec[0]-range,jVec[0]+range,iVec[0]-range,iVec[0]+range,5);
> fitsim1_5->SetParameters(pVec[0]/100,jVec[0],1.,iVec[0],1.);
>
> TF2 *fitsim2_5 =
> new
>TF2("fitsim2_5",Gaus2D,jVec[1]-range,jVec[1]+range,iVec[1]-range,iVec[1]+range,5);
> fitsim2_5->SetParameters(pVec[1]/100,jVec[1],1.,iVec[1],1.);
>
>and fit the histogram:
>
> histsim->Fit("fitsim1_5","RN");
> histsim->Fit("fitsim2_5","RN");
>
>And it works. However, I would like to create the fit functions in a
>loop and
>make the fitting in the same loop for several different regions, such
>as:
>
>const int isize = pVec.size(); //// isize = 2 since there are two peaks
>to
>fit here.
> TF2 *fitarr[isize];
> for (int is=0; is<isize; is++)
> {
> char afitn[20];
> int range = 10;
> sprintf(afitn, "fitsim3_%d", isize);
> double xmin = double(jVec[is] -range);
> double xmax = double(jVec[is] +range);
> double ymin = double(iVec[is] -range);
> double ymax = double(iVec[is] +range);
> double xctr = double(jVec[is]);
> double yctr = double(iVec[is]);
> double peak = double(pVec[is]/100);
>
> fitarr[is] = new TF2(afitn,Gaus2D,xmin,xmax,ymin,ymax,5);
> fitarr[is]->SetParameters(peak,xctr,1.,yctr,1.);
>
> fitarr[is]->SetLineColor(kBlack);
> histsim->Fit(afitn,"RN+");
> fitarr[is]->SetRange(0,0, 256,256);
> fitarr[is]->SetMaximum(histsim->GetMaximum());
> fitarr[is]->Draw("cont1, same");
> gPad->Modified();
> gPad->Update();
> }
>
>When I do this, it only plots the last fit in the loop. The first fit
>is lost
>somewhere and not plotted. If I do the loop only for one fit function
>(by
>changing isize to 1), it plots the first fit. When isize =2, it only
>plots the
>second fit. I don't see any difference between the first method (which
>works)
>and the second method (that fails). Unfortunately I am not able to send
>a
>working code since the input files are rather big. But the main idea is
>to do
>multiple fits on the same histogram (on different ranges of it) by
>using a for
>loop. I appreciate your comments if I am doing something wrong in the
>for
>loop.
>Thanks,
>
>- Nevzat Guler
>(757) 553-8318
Received on Thu Dec 02 2010 - 00:22:27 CET