Loop counter and log10 of a variable

[Federico_Versari]

Hi everyone,

I started working with TDataFrame but I have two questions:

1. Is there a way to print on the terminal some kind of iterator that show me the status of the event loop ? I’m running in multithreading
2. If I have a branch of my TTree called “Energy” and i want to plot the log10(Energy) in a TH1D how I can do it ?
   Because if I try something like:
   auto myHisto = d.Histo1D(“log10(Energy)”);
   The error is that “log10(Energy)” is not a branch of the TTree

(this is the first time that I write something in this forum so maybe I’m doing something not in the proper way)
Cheers,
Federico

[eguiraud]

Hi Federico,
you did everything the proper way, your question is very clear and already tagged as #tdataframe.

Printing status of the event loop
There are a few different ways to do this.
In C++, the easiest and most robust is via TResultProxy::OnPartialResult.
A TResultProxy is the pointer-like object that all of TDF actions return: for example, d.Histo1D returns a TResultProxy<TH1D>. OnPartialResult is a method that registers a function that is to be called on the contents of the TResultProxy every N processed entries.

Things should be clear with an example:

// print "hello" every 100 processed entries
auto h = tdf.Histo1D("x");
auto printHello = [](TH1D&) { std::cout << "hello" << std::endl; };
h.OnPartialResult(100, printHello);
h->GetNEntries(); // event loop is run here, when we use the result for the first time

As per the docs, in a multi-thread execution OnPartialResult is only called by one of the threads, so it will just give you a rough idea of how far the event loop went. You can also check $ROOTSYS/tutorials/dataframe/tdf013_InspectAnalysis.C for a nicer example usage of OnPartialResult, also in a multi-thread application – there we plot the result histogram and update it every 50 entries dynamically.

Histogram of the log10 of a variable
This is very easy, you need to define a new column with the logarithm of your energy variable:

auto hlog10 = tdf.Define("log10E", "log10(Energy)").Histo1D("log10E");

If you are using python instead answer number 1 is to be revised a bit.
Hope this helps!
Enrico

[Federico_Versari]

Thank you very much !!
This is exactly what I needed !
Federico