RDataFrame column not found

[Sergio_Gonzalez]

Hello All,

I am trying to use an RDataFrame object to get histograms and I use the following lines:

auto h = dataFrame.Filter(selection.Data())
                      .Define(w_name.Data() , str_weight.Data())
                      .Define(var_name.Data(), branch_name.Data())                    .Histo1D(ROOT::RDF::TH1DModel(*histogram.hist),var_name.Data(),w_name.Data());

After that, I want to add an extra selection to the filter and get another histogram. I do this like this:

auto h_perDSID = dataFrame.Filter(selection_perDSID.Data())
                                      .Histo1D(ROOT::RDF::TH1DModel(*histogram.perDSIDHists[dsid]), var_name.Data(), w_name.Data());

But when I do that I get these errors:

what(): Unknown columns: $$var_name.Data(), w_name.Data()$$

and if I try to Define them as with the first call I get a Redefinition error…

I’m a bit confused because it says that the column is not found but when I try to define them it says I’m trying to redefine it.

Anyone has some ideas?

Thanks!
Sergio

[eguiraud]

Hi,
from the error message it seems like RDF is trying to access columns called "var_name.Data()", so maybe you are passing certain arguments as strings (with commas, like in "var_name.Data()" instead of just var_name.Data()). That would be my guess.

Otherwise I think we are going to need a small reproducer of your problem that we can run: from the snippets you posted above, things seem fine.

Cheers,
Enrico