I need to divide two TGraphErrors. On googling, I came across an old post root.cern.ch/root/roottalk/roottalk04/2261.html
which addresses this issue. The prescription was to convert the TGraphErrors into histograms and use TGraphAsymmErrors:BayesDivide.
I would like to know if a new implementation of ROOT allows an easier way to do this, taking into account the correct error estimation.
there is no Divide function in ROOT for two TGraphErrors, since in general does not make much sense since one cannot be sure that the two graphs have the same abscissa (x) values.
In addition specific assumptions must be made on the type of errors for the two graphs.
(see the reply to the post indicated by you root.cern.ch/root/roottalk/roottalk04/2266.html )
So I also think it is better for the user to perform this operation by looping on the graph points