Hi Rene,
Indeed, there is a division by zero in my function when evaluated at the
lower bound. I guess I was confused as to what ROOT did when it saved a
TF1 object. Readjusting the lower bound slightly fixes everything.
But this brings up another issue, though. Certainly a TF1 object is
evaluated in the Draw() method. It seems inconsistent that Write()
complains (as it should, I guess) and Draw() gives no warnings.
Maybe Write() and Draw() decide to evaluate the function at differing sets
of points?
Anyway, thanks for your help. I appreciate it.
-Kalen
-----------------------------------------------------------------
On Wed, 21 May 2003, Rene Brun wrote:
> Hi Kalen,
>
> What you do should work. Could you send me a copy of your function?
>
> When writing a TF1 referencing an interpreted or compiled function,
> the function is tabulated at fNpx values in such a way that you can
> always draw the function in the reading program even if the code
> of the function is not available.
> In your case, it looks like the function is being evaluated with a
> division by 0.
>
> Rene Brun
>
> On Wed,
> 21 May 2003, Kalen Martens wrote:
>
> > Hi again,
> >
> > Is it possible to save a function (a TF1 object, for example) to a .root
> > file? I am having difficulties when I try:
> >
> > root [14] TFile f("DeltaNLR_functions.root","recreate");
> >
> > root [15] TF1 *qq_mh100 = new TF1("qq_mh100", DeltaNLR_qq_1D_mh100, 350.,
> > 1200., 0);
> >
> > ...define more functions...
> >
> > root [26] qq_mh100 -> Write();
> > Error: operator '/' divided by zero FILE:/tmp/sxAeOo_cint LINE:17
> > *** Interpreter error recovered ***
> >
> > It looks like the function is being evaluated. The Draw() method works
> > just fine and without any errors, but the Write() is doing something
> > strange.
> >
> > Thanks,
> >
> > -Kalen
> >
> > -----------------------------------------------------------------
> > - Kalen Martens Ph:(416)-978-6632 -
> > - Dept. of Physics -
> > - University of Toronto, Toronto, Canada -
> > -----------------------------------------------------------------
> >
>
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:11 MET