RDataFrame analysis tutorials

These examples show various features of RDataFrame: ROOT's declarative analysis interface.

RDataFrame offers a high level interface for the analysis of data stored in TTrees, CSV files and other data formats.

In addition, multi-threading and other low-level optimisations allow users to exploit all the resources available on their machines transparently.

In a nutshell:

ROOT::EnableImplicitMT(); // Enable ROOT's implicit multi-threading
ROOT::RDataFrame d("myTree", "file_*.root"); // Interface to TTree and TChain
auto histoA = d.Histo1D("Branch_A");  // Book the filling of a histogram
auto histoB = d.Histo1D("Branch_B");  // Book the filling of another histogram
// Data processing is triggered by the next line, which accesses a booked result for the first time
// All booked results are evaluated during the same parallel event loop.
histoA->Draw(); // <-- event loop runs here!
histoB->Draw(); // HistoB has already been filled, no event loop is run here

Explore the examples below or go to RDataFrame's user guide. A list of all the RDataFrame tutorials can be found here.

Table of contents

• Introduction
• Processing your data
• Write and read from many sources
• Interface with Numpy and Pandas
• Distributed execution in Python
• Know more about your analysis
• Example HEP analyses tutorials
• List of all the tutorials

Introduction

To get started these examples show how to create a simple RDataFrame, how to process the data in a simple analyses and how to plot distributions.

Tutorial		Description
df000_simple.C	df000_simple.py	Simple RDataFrame example in C++.
df001_introduction.C	df001_introduction.py	Basic RDataFrame usage.
df002_dataModel.C	df002_dataModel.py	Show how to work with non-flat data models, e.g. vectors of tracks.

Processing your data

A collection of building block examples for your analysis.

Tutorial		Description
df003_profiles.C	df003_profiles.py	Use TProfiles.
df005_fillAnyObject.C		Fill any object the class of which exposes a Fill method
df006_ranges.C	df006_ranges.py	Use Range to limit the amount of data processed.
df012_DefinesAndFiltersAsStrings.C	df012_DefinesAndFiltersAsStrings.py	Use just-in-time-compiled Filters and Defines for quick prototyping.
df016_vecOps.C	df016_vecOps.py	Process collections in RDataFrame with the help of RVec.
df018_customActions.C		Implement a custom action to fill THns.
df020_helpers.C		Show usage of RDataFrame's helper tools.
df021_createTGraph.C	df021_createTGraph.py	Fill a TGraph.
df022_useKahan.C		Implement a custom action that evaluates a Kahan sum.
df023_aggregate.C		Use the Aggregate action to specify arbitrary data aggregations.
df025_RNode.C		Manipulate RDF objects in functions, loops and conditional branches.
df036_missingBranches.C		Deal with missing values due to a missing branch when switching to a new file in a chain.
df037_TTreeEventMatching.C		Deal with missing values due to not finding a matching event in an auxiliary dataset.

Write and read from many sources

The content of a dataframe can be written to a ROOT file. In addition to ROOT files, other file formats can be read.

Tutorial		Description
df007_snapshot.C	df007_snapshot.py	Write out a dataset.
df008_createDataSetFromScratch.C	df008_createDataSetFromScratch.py	Generate data from scratch.
df009_FromScratchVSTTree.C		Compare creation of a ROOT dataset with RDataFrame and TTree.
df010_trivialDataSource.C	df010_trivialDataSource.py	Simplest possible data source.
df014_CSVDataSource.C	df014_CSVDataSource.py	Process a CSV.
df015_LazyDataSource.C		Concatenate computation graphs with the "lazy data source.
df019_Cache.C	df019_Cache.py	Cache a processed RDataFrame in memory for further usage.
df027_SQliteDependencyOverVersion.C		Analyse a remote sqlite3 file.
df028_SQliteIPLocation.C		Plot the location of ROOT downloads reading a remote sqlite3 file.
df029_SQlitePlatformDistribution.C		Analyse data in a sqlite3 file.
df030_SQliteVersionsOfROOT.C		Analyse data in a sqlite3 file and create a plot.

Interface with Numpy and Pandas

From Python, NumPy arrays can be imported into RDataFrame and columns from RDataFrame can be converted to NumPy arrays. A Pandas DataFrame can also be converted into a RDataFrame.

Tutorial	Description
df026_AsNumpyArrays.py	Read data into Numpy arrays.
df032_RDFFromNumpy.py	Read data from Numpy arrays.
df035_RDFFromPandas.py	Read data from Pandas DataFrame.

Distributed execution in Python

RDataFrame applications can be executed in parallel through distributed computing frameworks on a set of remote machines via Apache Spark or Dask.

Tutorial	Description
distrdf001_spark_connection.py	Configure a Spark connection and fill two histograms distributedly.
distrdf002_dask_connection.py	Configure a Dask connection and fill two histograms distributedly.
distrdf003_live_visualization.py	Configure a Dask connection and visualize the filling of a 1D and 2D histograms distributedly.

Know more about your analysis

In RDataFrame there exist methods to inspect the data and the computation graph.

Tutorial		Description
df004_cutFlowReport.C	df004_cutFlowReport.py	Display cut/Filter efficiencies.
df013_InspectAnalysis.C		Use callbacks to update a plot and a progress bar during the event loop.
df024_Display.C	df024_Display.py	Use the Display action to inspect entry values.
df031_Stats.C	df031_Stats.py	Use the Stats action to extract the statistics of a column.
	df033_Describe.py	Get information about your analysis.
df034_SaveGraph.C	df034_SaveGraph.py	Look at the DAG of your analysis

Example HEP analyses tutorials

With RDataFrame advanced analyses can be executed on large amounts of data. These examples shows how particle physics analyses can be carried out using Open Data from different experiments.

Tutorial		Description
df017_vecOpsHEP.C	df017_vecOpsHEP.py	Use RVecs to plot the transverse momentum of selected particles.
df101_h1Analysis.C		Express ROOT's standard H1 analysis.
df102_NanoAODDimuonAnalysis.C	df102_NanoAODDimuonAnalysis.py	Process NanoAOD files.
df103_NanoAODHiggsAnalysis.C	df103_NanoAODHiggsAnalysis.py	An example of complex analysis: reconstructing the Higgs boson.
	df104_HiggsToTwoPhotons.py	The Higgs to two photons analysis from the ATLAS Open Data 2020 release.
	df105_WBosonAnalysis.py	The W boson mass analysis from the ATLAS Open Data release of 2020.
df106_HiggsToFourLeptons.C	df106_HiggsToFourLeptons.py	The Higgs to four lepton analysis from the ATLAS Open Data release of 2020.
	df107_SingleTopAnalysis.py	A single top analysis using the ATLAS Open Data release of 2020.

Files
file	df000_simple.C
	Simple RDataFrame example in C++.
file	df000_simple.py
	Simple RDataFrame example in Python.
file	df001_introduction.C
	Basic RDataFrame usage.
file	df001_introduction.py
	Basic usage of RDataFrame from python.
file	df002_dataModel.C
	Show how to work with non-flat data models, e.g.
file	df002_dataModel.py
	Show how to work with non-flat data models, e.g.
file	df003_profiles.C
	Use TProfiles with RDataFrame.
file	df003_profiles.py
	Use TProfiles with RDataFrame.
file	df004_cutFlowReport.C
	Display cut/Filter efficiencies with RDataFrame.
file	df004_cutFlowReport.py
	Display cut/Filter efficiencies with RDataFrame.
file	df005_fillAnyObject.C
	Using the generic Fill action.
file	df006_ranges.C
	Use Range to limit the amount of data processed.
file	df006_ranges.py
	Use Range to limit the amount of data processed.
file	df007_snapshot.C
	Write ROOT data with RDataFrame.
file	df007_snapshot.py
	Write ROOT data with RDataFrame.
file	df008_createDataSetFromScratch.C
	Create data from scratch with RDataFrame.
file	df008_createDataSetFromScratch.py
	Create data from scratch with RDataFrame.
file	df009_FromScratchVSTTree.C
	Compare creation of a ROOT dataset with RDataFrame and TTree.
file	df010_trivialDataSource.C
	Use the "trivial data source", an example data source implementation.
file	df010_trivialDataSource.py
	Use the "trivial data source", an example data source implementation.
file	df012_DefinesAndFiltersAsStrings.C
	Use just-in-time-compiled Filters and Defines for quick prototyping.
file	df012_DefinesAndFiltersAsStrings.py
	Use just-in-time-compiled Filters and Defines for quick prototyping.
file	df013_InspectAnalysis.C
	Use callbacks to update a plot and a progress bar during the event loop.
file	df014_CSVDataSource.C
	Process a CSV file with RDataFrame and the CSV data source.
file	df014_CSVDataSource.py
	Process a CSV file with RDataFrame and the CSV data source.
file	df015_LazyDataSource.C
	Use the lazy RDataFrame data source to concatenate computation graphs.
file	df016_vecOps.C
	Process collections in RDataFrame with the help of RVec.
file	df016_vecOps.py
	Process collections in RDataFrame with the help of RVec.
file	df017_vecOpsHEP.C
	Use RVecs to plot the transverse momentum of selected particles.
file	df017_vecOpsHEP.py
	Use RVecs to plot the transverse momentum of selected particles.
file	df018_customActions.C
	Implement a custom action to fill THns.
file	df019_Cache.C
	Cache a processed RDataFrame in memory for further usage.
file	df019_Cache.py
	Cache a processed RDataFrame in memory for further usage.
file	df020_helpers.C
	Show usage of RDataFrame's helper tools, contained in ROOT/RDFHelpers.hxx.
file	df021_createTGraph.C
	Fill a TGraph using RDataFrame.
file	df021_createTGraph.py
	Fill a TGraph using RDataFrame.
file	df022_useKahan.C
	Implement a custom action that evaluates a Kahan sum.
file	df023_aggregate.C
	Use the Aggregate action to specify arbitrary data aggregations.
file	df024_Display.C
	Use the Display action to inspect entry values.
file	df024_Display.py
	Use the Display action to inspect entry values.
file	df025_RNode.C
	Manipulate RDF objects in functions, loops and conditional branches.
file	df026_AsNumpyArrays.py
	Read data from RDataFrame into Numpy arrays.
file	df027_SQliteDependencyOverVersion.C
	Plot the ROOT downloads based on the version reading a remote sqlite3 file.
file	df028_SQliteIPLocation.C
	Plot the location of ROOT downloads reading a remote sqlite3 file.
file	df029_SQlitePlatformDistribution.C
	Use RDataFrame to display data about ROOT downloads.
file	df030_SQliteVersionsOfROOT.C
	Read an sqlite3 databases with RDataFrame and plot statistics on ROOT downloads.
file	df031_Stats.C
	Use the Stats action to extract the statistics of a column.
file	df031_Stats.py
	Use the Stats action to extract the statistics of a column.
file	df032_RDFFromNumpy.py
	Read data from Numpy arrays into RDataFrame.
file	df033_Describe.py
	Get information about the dataframe with the convenience method Describe.
file	df034_SaveGraph.C
	Basic SaveGraph usage.
file	df034_SaveGraph.py
	Basic SaveGraph usage.
file	df035_RDFFromPandas.py
	Read data from Pandas Data Frame into RDataFrame.
file	df036_missingBranches.C
file	df037_TTreeEventMatching.C
file	df038_NumbaDeclare.py
	This tutorial illustrates how PyROOT supports declaring C++ callables from Python callables making them, for example, usable with RDataFrame.
file	df101_h1Analysis.C
	Show how to express ROOT's standard H1 analysis with RDataFrame.
file	df102_NanoAODDimuonAnalysis.C
	Show how NanoAOD files can be processed with RDataFrame.
file	df102_NanoAODDimuonAnalysis.py
	Show how NanoAOD files can be processed with RDataFrame.
file	df103_NanoAODHiggsAnalysis.C
	An example of complex analysis with RDataFrame: reconstructing the Higgs boson.
file	df103_NanoAODHiggsAnalysis.py
	An example of complex analysis with RDataFrame: reconstructing the Higgs boson.
file	df103_NanoAODHiggsAnalysis_python.h
	Header file with functions needed to execute the Python version of the NanoAOD Higgs tutorial.
file	df104_HiggsToTwoPhotons.py
	The Higgs to two photons analysis from the ATLAS Open Data 2020 release, with RDataFrame.
file	df105_WBosonAnalysis.py
	The W boson mass analysis from the ATLAS Open Data release of 2020, with RDataFrame.
file	df106_HiggsToFourLeptons.C
	The Higgs to four lepton analysis from the ATLAS Open Data release of 2020, with RDataFrame.
file	df106_HiggsToFourLeptons.py
	The Higgs to four lepton analysis from the ATLAS Open Data release of 2020, with RDataFrame.
file	df107_SingleTopAnalysis.py
	A single top analysis using the ATLAS Open Data release of 2020, with RDataFrame.
file	distrdf001_spark_connection.py
	Configure a Spark connection and fill two histograms distributedly.
file	distrdf002_dask_connection.py
	Configure a Dask connection and fill two histograms distributedly.
file	distrdf003_live_visualization.py
	Configure a Dask connection and visualize the filling of a 1D and 2D histograms distributedly.