Unbinlikelihood fit using event by event error

Hello I am Cecilia. I realy need some help. I am trying to perform an
unbinlikeliehood using Roofit. I have defined a rooargset with a given
variable and its uncertainty in the measure defined for each individual
event that comes from values in a tree.

I want to use the uncertainty in the measure to perform the fit. I tried
with doing:

exp_dec.fitTo(*DataSet, ConditionalObservables(error) ,SumW2Error(kTRUE)
,Range(start,end));

where exp_dec is an exponential roofit distribution called RooExponential, defined for a certain variable, and “error” its the variable uncertainty. Both are rooreal vars related into a rooargset.

I tried the same without the option ConditionalObserbable()

Also I tried writing: in the option RooFit::ConditionalObservables()

The result of the fit parameter and its uncertainty does not changes when
I used the option, so I think that is not working and I m doing something
wrong.

I tried reading different manuals like RooFit_Users_Manual_2.07-29.pdf
where I found the option and doing some of the tutorials like:

rf307_fullpereventerrors.c

rf306_condpereventerrors.c

, but in the examples distribution are used to describe the uncertainty. I
need to use the event by event error and I don’t understand how.

How can I perform a unbinlikelihoodbin using event by event uncertainty
that comes from a tree in a root file?
Is there another example that i can use?

Thank, you