RDataframe friend tree columns in Histo1D

[FoxWise]

I have a RDataFrame consisted of 2 trees like this.

ch = ROOT.TChain("tree1")
ch.Add("/path/to/file1.root")

ch1 = ROOT.TChain("tree2")
ch1.Add("/path/to/file2.root")

ch.AddFriend(ch1,  "tr2")
df = ROOT.RDataFrame(ch)

I can use columns of the friend tree fine with Filter and Define.
But to make them work in Histo1D i need create another temporary column - copy. Is it how its supposed to be? Or am I missing something?

# **doesn't work!**
h2 = df.Histo1D(("h2", "title", 10, 0, 10), "tr2.x")

# works fine
h3 = df.Define("tree2_x", "tr2.x")\
       .Histo1D(("h3", "title", 10, 0, 10), "tree2_x")

ROOT Version: 6.22
Platform: Centos 7
Compiler: g++ 9.2

---

[eguiraud]

Hi @FoxWise,
thanks for the report, it looks like a bug but I can’t reproduce it in v6.22/06. Does the following program fail for you?

import ROOT

if __name__ == "__main__":
  ROOT.RDataFrame(1).Define("x", "42").Snapshot("t", "f1.root")
  ROOT.RDataFrame(1).Define("x", "42").Snapshot("t", "f2.root")

  ch1 = ROOT.TChain("t")
  ch1.Add("f1.root")
  ch2 = ROOT.TChain("t")
  ch2.Add("f2.root")
  ch1.AddFriend(ch2, "tr2")

  df = ROOT.RDataFrame(ch1)
  print(df.Histo1D("tr2.x").GetMean())

If yes, does this also fail (needs to be compiled e.g. with g++ main.cpp $(root-config --libs --cflags)?

#include <ROOT/RDataFrame.hxx>
#include <TChain.h>
#include <iostream>

int main() {
  {
    TFile f("f.root", "recreate");
    TTree t("t", "t");
    int x = 42;
    t.Branch("x", &x);
    t.Fill();
    t.Write();
  }

  TChain ch1("t");
  ch1.Add("f.root");
  TChain ch2("t");
  ch2.Add("f.root");
  ch1.AddFriend(&ch2, "tr2");

  ROOT::RDataFrame df(ch1);
  std::cout << df.Histo1D("tr2.x")->GetMean() << '\n';
  return 0;
}

If any of these fail for you, can you please try with 6.22/06?
If not, what am I doing differently?

Cheers,
Enrico

[FoxWise]

Hi @eguiraud
thanks for the fast feedback. Sorry, I oversimplified my case in the first place.

Here is the code I would like to work without re-defining thingie:

import ROOT

vectors = '''
#include "Math/Vector3D.h"
using namespace ROOT::Math;
'''

ROOT.gInterpreter.Declare(vectors)

if __name__ == "__main__":
    ROOT.RDataFrame(1).Define("vec", "XYZVector(42, 42*42, 42*42*42)").Snapshot("t", "f1.root")
    ROOT.RDataFrame(1).Define("vec", "XYZVector(42, 42*42, 42*42*42)").Snapshot("t", "f2.root")

    ch1 = ROOT.TChain("t")
    ch1.Add("f1.root")
    ch2 = ROOT.TChain("t")
    ch2.Add("f2.root")
    ch1.AddFriend(ch2, "tr2")

    df = ROOT.RDataFrame(ch1)
    df = df.Define("tr2_vec_Z", "tr2.vec.Z()")
    print(df.Histo1D("tr2_vec_Z").GetMean()) # works
    print(df.Histo1D("tr2.vec.Z()").GetMean()) # doesn't work

But probably doing calculations/executing methods are not allowed inside Histo1D? Is there no more fancy way?

cheers,
Bohdan

[eguiraud]

Ah alright
Histo1D does not accept expressions, just column names. Define is the right thing to use in this case. It’s a bit more verbose but it keeps responsibilities clearly separated.

Cheers,
Enrico