Institut fr Experimentelle Kernphysik

A ROOT Guide For Students

Diving Into ROOT

Abstract:

ROOT is an object-oriented framework for data analysis. Among its prominent features are an advanced
graphical user interface for visualization and interactive data analysis and an interpreter for the C++
programming language, which allows rapid prototyping of analysis code based on the C++ classes provided by ROOT. Access to ROOT classes is also possible from the very versatile and popular scripting
language

Python.

This introductory guide shows the main features applicable to typical problems of data analysis in student labs: input and plotting of data from measurements and comparison with and tting of analytical
functions. Although appearing to be quite a heavy

http://root.cern.ch

gun for some of the simpler problems, getting used to

a tool like ROOT at this stage is an optimal preparation for the demanding tasks in state-of-the art,
scientic data analysis.

Authors:
Danilo
Gnter

Piparo
Quast
Zeise

,
,

Manuel

Version of July 27, 2015

CHAPTER 1
MOTIVATION AND INTRODUCTION

Welcome to data analysis !

Comparison of measurements to theoretical models is one of the standard tasks in experimental physics.
In the most simple case, a model is just a function providing predictions of measured data. Very often,
the model depends on parameters. Such a model may simply state the current
voltage

U ,

is proportional to the

and the task of the experimentalist consists of determining the resistance,

R,

from a set of

measurements.
As a rst step, a visualisation of the data is needed. Next, some manipulations typically have to be
applied, e. g. corrections or parameter transformations. Quite often, these manipulations are complex
ones, and a powerful library of mathematical functions and procedures should be provided - think for
example of an integral or peak-search or a Fourier transformation applied to an input spectrum to obtain
the actual measurement described by the model.
One specialty of experimental physics are the inevitable errors aecting each measurement, and visualization tools have to include these. In subsequent analysis, the statistical nature of the errors must be
handled properly.
As the last step, measurements are compared to models, and free model parameters need to be determined in this process , see Figure1.1 for an example of a function (model) t to data points. Several
standard methods are available, and a data analysis tool should provide easy access to more than one of
them. Means to quantify the level of agreement between measurements and model must also be available.

1
0.9
0.8

Y = f (x)

0.7
0.6
0.5
0.4

Data

0.3

Model
0.2
0

Fi

oo

R
b
La

Figure 1.1.:

Measured data points with error bars and tted quadratic function .

1. Motivation and Introduction

Quite often, the data volume to be analyzed is large - think of ne-granular measurements accumulated
with the aid of computers. A usable tool therefore must contain easy-to-use and ecient methods for
data handling.
In Quantum mechanics, models typically only predict the probability density function (pdf ) of measurements depending on a number of parameters, and the aim of the experimental analysis is to extract
the parameters from the observed distribution of frequencies at which certain values of the measurement
are observed. Measurements of this kind require means to generate and visualize frequency distributions,
so-called histograms, and stringent statistical treatment to extract the model parameters from purely
statistical distributions.
Simulation of expected data is another important aspect in data analysis. By repeated generation of
pseudo-data, which are analysed in the same manner as intended for the real data, analysis procedures
can be validated or compared. In many cases, the distribution of the measurement errors is not precisely
known, and simulation oers the possibility to test the eects of dierent assumptions.

1.1. Welcome to ROOT

A powerful software framework addressing all of the above requirements is ROOT [1], an open source
project coordinated by the European Centre for Particle Physics, CERN in Geneva. ROOT is very exible
and provides both a programming interface to use in own applications and a graphical user interface for
interactive data analysis. The purpose of this document is to serve as a beginners guide and provides
extendable examples for your own use cases, based on typical problems addressed in student labs. This
guide will hopefully lay the ground for more complex applications in your future scientic work building
on a modern, state-of the art tool for data analysis.
This guide in form of a tutorial is intended to introduce you to the ROOT package in about

50

pages.

This goal will be accomplished using concrete examples, according to the learning by doing principle.
Also because of this reason, this guide cannot cover the complexity of the ROOT package. Nevertheless,
once you feel condent with the concepts presented in the following chapters, you will be able to appreciate
the ROOT Users Guide [2] and navigate through the Class Reference [3] to nd all the details you might
be interested in. You can even look at the code itself, since ROOT is a free, open-source product. Use
these documents in parallel to this tutorial!
The ROOT Data Analysis Framework itself is written in and heavily relys on the programming language

C++, and therefore some knowledge about C andC++ is required. Eventually, just prot from the immense
available literature about C++ if you do not have any idea of what object oriented programming could be.
Recently, an alternative and very powerful way to use and control ROOT classes via the interpreted
high-level programming language

Python became available. Python itself oers powerful modules and

packages for data handling, numerical applications and scienc computing. A vast number of bindings
or wrappers to packages and tools written in other languages is also available.
functionality is provided by the ROOT package
based on

PyRoot

Access to the ROOT

[5], allowing interactive work as well as scritps

Python. This is presented at the end of this guide in Chapter 8.

ROOT is available for many platforms (Linux, Mac OS X, Windows. . . ), but in this guide we will
implicitly assume that you are using Linux. The rst thing you need to do with ROOT is install it. Or do
you? Obtaining the latest ROOT version is straightforward. Just seek the Pro version on this webpage

http://root.cern.ch/drupal/content/downloading-root.

You will nd precompiled versions for the

dierent architectures, or the ROOT source code to compile yourself. Just pick up the avour you need
and follow the installation instructions. Or even simpler: use a virtual machine with ROOT installed
ready for use, as availalbe under e. g.

http://www-ekp.physik.uni-karlsruhe.de/~quast.

Let's dive into ROOT!

CHAPTER 2
ROOT BASICS

Now that you have installed ROOT, what's this interactive shell thing you're running?
ROOT leads a double life.

It's like this:

It has an interpreter for macros (CINT [4]) that you can run from the

command line or run like applications.

But it is also an interactive shell that can evaluate arbitrary

statements and expressions. This is extremely useful for debugging, quick hacking and testing. Let us
rst have a look at some very simple examples.

2.1. ROOT as calculator

You can even use the ROOT interactive shell in lieu of a calculator! Launch the ROOT interactive shell
with the command
1

>

root

on your Linux box. The prompt should appear shortly:

1

root

[1]

and let's dive in with the steps shown here:

1
2
3
4
5
6
7
8
9
10
11
12

root [ 0 ] 1+1
( const int ) 2
root [ 1 ] 2 * ( 4 + 2 ) / 1 2 .
( const double ) 1 . 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e +00
root [ 2 ] sqrt ( 3 )
( const double ) 1 . 7 3 2 0 5 0 8 0 7 5 6 8 8 7 7 1 9 e +00
root [ 3 ] 1 > 2
( const int ) 0
root [ 4 ] TMath : : Pi ( )
( Double_t ) 3 . 1 4 1 5 9 2 6 5 3 5 8 9 7 9 3 1 2 e +00
root [ 5 ] TMath : : Erf ( . 2 )
( Double_t ) 2 . 2 2 7 0 2 5 8 9 2 1 0 4 7 8 4 4 7 e 01
Not bad. You can see that ROOT oers you the possibility not only to type in
advanced mathematical functions, which live in the

TMath

C++

statements, but also

namespace.

Now let's do something more elaborated. A numerical example with the well known geometrical series:
1
2
3
4
5
6

root [ 6 ] double x =.5

root [ 7 ] int N =30
root [ 8 ] double geom_series =0
root [ 9 ] for ( int i =0; i<N ;++ i ) geom_series+=TMath : : Power ( x , i )
root [ 1 0 ] TMath : : Abs ( geom_series (1 TMath : : Power ( x , N 1) ) /(1 x ) )
( Double_t ) 1 . 8 6 2 6 4 5 1 4 9 2 3 0 9 5 7 0 3 e 09

2. ROOT Basics

Here we made a step forward. We even declared variables and used a

for

control structure. Note that

there are some subtle dierences between CINT and the standard C++ language. You do not need the
"; at the end of line in interactive mode  try the dierence e.g. using the command at line

root [6].

2.2. ROOT as Function Plotter

Using one of ROOT's powerful classes, here

1 , will allow us to display a function of one variable,

TF1

x.

Try the following:

1
2

root
root

f1

[11]
[12]

TF1 * f1 = new TF1 ( " f1 " , " sin (x)/x"

f1>Draw ( ) ;

,0. ,10.) ;

is a pointer to an instance of a TF1 class, the arguments are used in the constructor; the rst one

of type string is a name to be entered in the internal ROOT memory management system, the second

sin(x)/x, and the two parameters of type real dene

Draw() method, here without any parameters, displays the function in a

string type parameter denes the function, here

the range of the variable

x.

The

window which should pop up after you typed the above two lines. Note again dierences between CINT
and C++: you could have omitted the "; at the end of lines, of CINT woud have accepted the "." to
access the method

Draw().

However, it is best to stick to standard C++ syntax and avoid CINT-specic

code, as will become clear in a moment.

A slightly extended version of this example is the denition of a function with parameters, called [0],
[1] and so on in ROOT formula syntax. We now need a way to assign values to these parameters; this is
achieved with the method

SetParameter(<parameter_number>,<parameter_value>) of class TF1.

Here

is an example:
1
2
3
4

root
root
root
root

[13]
[14]
[15]
[16]

TF1 * f1 = new TF1 ( " f2 " , " [0]* sin ([1]* x)/x"
f1>SetParameter ( 0 , 1 ) ;
f1>SetParameter ( 1 , 1 ) ;
f1>Draw ( ) ;

,0. ,10.) ;

Of course, this version shows the same results as the initial one. Try playing with the parameters and plot

TF1

the function again. The class

has a large number of very useful methods, including integration and

dierentiation. To make full use of this and other ROOT classes, visit the documentation on the Internet
under

http://root.cern.ch/drupal/content/reference-guide. Formulae in ROOT are evaluated

TFormula, so also look up the relevant class documentation for examples, implemented

using the class

functions and syntax.

On many systems, this class reference-guide is available locally, and you should denitely download it
to your own system to have it at you disposal whenever you need it.
To extend a little bit on the above example, consider a more complex function you would like to dene.
You can also do this using standard

or

C++

code. In many cases this is the only practical way, as the

ROOT formula interpreter has clear limitations concerning complexity and speed of evaluation.
Consider the example below, which calculates and displays the interference pattern produced by light
falling on a multiple slit. Please do not type in the example below at the ROOT command line, there is
a much simpler way: Make sure you have the le

slits.cxx on disk, and type root slits.cxx in the

slit.cxx, i. e. all the lines in the le will be

shell. This will start root and make it read the macro
executed one after the other.

example

to

draw

falling

on

and

/*

***

ratio

the

grid
of

interference

with

slit

widht

pattern

of

light

slits
over

distance

between

4
5
6
7
8

/*

function

code

i n C */

double single ( double * x , double * par ) {

double const pi =4* atan ( 1 . ) ;
return pow ( sin ( pi * par [ 0 ] * x [ 0 ] ) / ( pi * par [ 0 ] * x [ 0 ] )

,2) ;

9
10
11
12

double nslit0 ( double * x , double * par ) {

double const pi =4* atan ( 1 . ) ;
return pow ( sin ( pi * par [ 1 ] * x [ 0 ] ) / sin ( pi * x [ 0 ] )
1 All

ROOT classes start with the letter T.

,2) ;

slits

***

*/

2.3. Controlling ROOT

13
14
15

double nslit ( double * x , double * par ) {

return single ( x , par ) * nslit0 ( x , par ) ;

16
17
18
19

/*

This

is

the

void slits ( )
float r , ns ;

main

program

*/

20
21

/*

request

user

input

*/

cout << " slit width / g ? " ;

scanf ( "%f" ,& r ) ;
cout << "# of slits ? " ;
scanf ( "%f" ,& ns ) ;
cout <<" interference pattern for "<< ns<<" slits , width / distance : "<<r<<endl ;

22
23
24
25
26
27
28
29
30

/* d e f i n e

function

and

set

options

*/

TF1 * Fnslit = new TF1 ( " Fnslit " , nslit , 5 . 0 0 1 , 5 . , 2 ) ;

Fnslit >SetNpx ( 5 0 0 ) ;
/ / s e t number o f p o i n t s

to

500

31

Fnslit >SetParameter ( 0 , r ) ;
Fnslit >SetParameter ( 1 , ns ) ;

32
33

// s e t

parameters ,

as

read

in

above

34

Fnslit >Draw ( ) ;

35
36

//

draw

the

interference

pattern

for

grid

with

slits

}
le:

slits.cxx

The example rst asks for user input,

namely the ratio of slit width over slit distance, and the number of slits.

nslit

After en-

tering this information, you should see the

graphical output as is shown in Figure 2.1

3.5

below.
This

is

more

complicated

example

than the ones we have seen before,

2.5

so

spend some time analysing it carefully,

you

1.5

should

continuing.

have

understood

it

before

Let us go through in de-

tail:

1
0.5
0

Lines 6-19 dene the necessary functions

-4

-2

in

C++

code, split into three separate func-

tions, as suggested by the problem consid-

Figure 2.1.: Output of macro slits.cxx with parameters 0.2 and

2.

ered. The full interference pattern is given

by the product of a function depending on
the ratio of the width and distance of the
slits, and a second one depending on the
number of slits. More important for us here

is the denition of the interface of these functions to make them usable for the ROOT class
is the pointer to

x,

the second one points to the array of parameters.

The main program starts in line 17 with the denition of a function

slits()

of type

TF1:

the rst argument

void.

After asking for

user input, a ROOT function is dened using the C-type function given in the beginning. We can now use all
methods of the

TF1

class to control the behaviour of our function  nice, isn't it?

If you like, you can easily extend the example to also plot the interference pattern of a single slit, using function

double single,

or of a grid with narrow slits, function

double nslit0,

in

TF1

instances.

Here, we used a macro, some sort of lightweight program, that the interpreter distributed with ROOT, CINT,
is able to execute. This is a rather extraordinary situation, since C++ is not natively an interpreted language!
There is much more to say, therefore there is a dedicated chapter on macros.

2.3. Controlling ROOT

One more remark at this point: as every command you type into ROOT is usually interpreted by CINT, an
escape character is needed to pass commands to ROOT directly. This character is the dot at the beginning of

2. ROOT Basics

a line:
1

root

. < command >

[1]

To

quit root, simply type .q

obtain a list of commands, use .?
access the shell of the operating system, type .!<OS_command>; try, e. g. .!ls or .!pwd
execute a macro, enter .x <file_name>; in the above example, you might have used .x slits.cxx

at

the ROOT prompt

load a macro,

.L slits.cxx

type

.L <file_name>;

in the above example, you might instead have used the command

followed by the function call

slits();.

Note that after loading a macro all functions and

procedures dened therein are available at the ROOT prompt.

2.4. Plotting Measurements

TGrapErrors with
ExampleData.txt in text format:

To display measurements in ROOT, including errors, there exists a powerful class

types of constructors. In the example here, we use data from the le
1
2

root
root

dierent

TGraphErrors * gr=new TGraphErrors ( " ExampleData . txt " ) ;

gr>Draw ( " AP " ) ;

[0]
[1]

You should see the output shown in Figure 2.2.

Make sure the le

ExampleData.txt

is

available in the directory from which you

started ROOT. Inspect this le now with
your favourate editor, or use the command

less ExampleData.txt

to inspect the le,

you will see that the format is very simple and easy to understand.
ning with

Lines begin-

# are ignored, very convenient to

add some comments on the type of data.

The data itself consist of lines with four
real numbers each, representing the x- and
y- coordinates and their errors of each data
point. You should quit

Draw("AP")
TGraphPainter

The argument of the method

is important here. It tells the

class to show the axes and to plot markers at the

x and y

positions of the specied

data points. Note that this simple example

Figure 2.2.:

Visualisation of data points with errors using the

class TGraphErrors

relies on the default settings of ROOT, concerning the size of the canvas holding the
plot, the marker type and the line colours
and thickness used and so on.

In a well-

written, complete example, all this would

need to be specied explicitly in order to obtain nice and reproducible results.
explain many more of the features of the class

TGraphErrors

A full chapter on graphs will

and its relation to other ROOT classes in much

more detail.

2.5. Histograms in ROOT

Frequency distributions in ROOT are handled by a set of classes derived from the histogram class
case

TH1F.

The letter

stands for "oat", meaning that the data type

float

histogram bin.

1
2
3
4
5
6

root
root
root
root
root
root

[0]
[1]
[2]
[3]
[4]
[5]

TH1,

in our

is used to store the entries in one

TF1 efunc ( " efunc " , " exp ([0]+[1]* x)" , 0 . , 5 . ) ;

efunc . SetParameter ( 0 , 1 ) ;
efunc . SetParameter ( 1 , 1 ) ;
TH1F * h=new TH1F ( "h" , " example histogram " , 1 0 0 , 0 . , 5 . ) ;
for ( int i =0; i < 1 0 0 0 ; i++) { h>Fill ( efunc . GetRandom ( ) ) ; }
h>Draw ( ) ;

2.6. Interactive ROOT

The rst three lines of this example dene a function, an exponential in this case, and set its parameters. In
Line 4 a histogram is instantiated, with a name, a title, a certain number of 100 bins (i. e. equidistant, equally
sized intervals) in the range from 0. to 5.
We

use

yet

another

new

feature

of

ROOT to ll this histogram with data,

example histogram

namely pseudo-random numbers generated

h
Entries
Mean
RMS

50

1000
0.9719
0.927

with the method

TF1::GetRandom, which in

turn uses an instance of the ROOT class

TRandom

40

created when ROOT is started.

Data is entered in the histogram in line

5 using the method

30

construct.

TH1F::Fill

As a result,

in a loop

the histogram

is lled with 1000 random numbers dis20

tributed according to the dened function. The histogram is displayed using the

10

0
0

method

TH1F::Draw().

You may think of

this example as repeated measurements of

0.5

1.5

2.5

3.5

4.5

the life time of a quantum mechanical state,

which are entered into the histogram, thus

Figure 2.3.:

Visualisation of a histogram lled with exponentially distributed, random numbers.

giving a visual impression of the probability density distribution. The plot is shown
in Figure 2.3.
Note that you will not obtain an identical plot when executing the above lines,

depending on how the random number generator is initialised.

The class

TH1F

does not contain a convenient input format from plain text les. The following lines of C++

code do the job. One number per line stored in the text le expo.dat is read in via an input stream and lled
in the histogram until end of le is reached.

1
2
3
4
5
6

root
root
root
root
root
root

[1]
[2]
[3]
[4]
[5]
[6]

TH1F * h=new TH1F ( "h" , " example histogram "

ifstream inp ; double x ;
inp . open ( " expo . dat " ) ;
while ( ! ( inp >> x ) ==0){ h>Fill ( x ) ; }
h>Draw ( ) ;
inp . close ( ) ;

,100 ,0. ,5.) ;

Histograms and random numbers are very important tools in statistical data analysis, and the whole Chapter 5
will be dedicated to this.

2.6. Interactive ROOT

Look at one of your plots again and move the mouse across. You will notice that this is much more than a static
picture, as the mouse pointer changes its shape when touching objects on the plot.

When the mouse is over

an object, a right-click opens a pull-down menu displaying in the top line the name of the ROOT class you are
dealing with, e.g.

TPaveText for
TGraphErrors

TCanvas

for the display window itself,

TFrame

for the frame of the plot,

TAxis

for the axes,

the plot name. Depending on which plot you are investigating, menus for the ROOT classes
or

TH1F

TF1,

will show up when a right-click is performed on the respective graphical representations.

The menu items allow direct access to the members of the various classes, and you can even modify them, e.g.
change colour and size of the axis ticks or labels, the function lines, marker types and so on. Try it!
You will probably like the following:
in the output produced by the example

slits.cxx,

right-click on the function line

and select "SetLineAttributes", then leftclick on "Set Parameters". This gives access to a panel allowing you to interactively
change the parameters of the function, as
shown in Figure 2.4. Change the slit width,
or go from one to two and then three or

Figure 2.4.:

Interactive
parameters.

ROOT

panel

for

setting

function

more slits, just as you like. When clicking

on "Apply", the function plot is updated
to reect the actual value of the parameters you have set.

2. ROOT Basics

Another very useful interactive tool is the

for the classes

TGraphErrors

and

TH1F.

FitPanel,

available

Predened t functions

can be selected from a pull-down menu, including  gaus,  expo

and  pol0 -  pol9 for Gaussian and exponential functions or

polynomials of degree 0 to 9, respectively.

In addition, user-

dened functions using the same syntax as for functions with parameters are possible.
After setting the initial parameters, a t of the selected function to the data of a graph or histogram can be performed and the
result displayed on the plot. The t panel is shown in Figure 2.5.
The t panel has a large number of control options to select the
t method, x or release individual paramters in the t, to steer
the level of output printed on the console, or to extract and display additional information like contour lines showing parameter
correlations.

Most of the methods of the class

TVirtualFitter

are easily available through the latest version of the graphical interface. As function tting is of prime importance in any kind of
data analysis, this topic will again show up in later chapters.
If you are satised with your plot, you probably want to save
it. Just close all selector boxes you opened previously, and select
the menu item Save as from the menu line of the window, which
will pop up a le selector box to allow you to choose the format,
le name and target directory to store the image.
There is one very noticeable feature here: you can store a plot
as a root macro! In this macro, you nd the C++ representation
of all methods and classes involved in generating the plot. This is
a very valuable source of information for your own macros, which
you will hopefully write after having worked through this tutorial.
Using the interactive capabilities of ROOT is very useful for
a rst exploration of possibilities. Other ROOT classes you will
be encountering in this tutorial have such graphical interfaces as

Figure 2.5.:

well. We will not comment further on this, just be aware of the

Fit functions to graphs and

histograms.

existence of interactive features in ROOT and use them if you nd

convenient. Some trial-and-error is certainly necessary to nd your way through the enormous number of menus
and possible parameter settings.

2.7. ROOT Beginners' FAQ

At this point of the guide, some basic question could have already come to your mind. We will try to clarify some
of them with further explanations in the following.

2.7.1. ROOT type declarations for basic data types

In the ocial ROOT documentation, you nd special data types replacing the normal ones, e. g.

Float_t

or

Int_t

replacing the standard

double, float

or

int

Double_t,

types. Using the ROOT types makes it easier to

port code between platforms (64/32 bit) or operating systems (windows/Linux), as these types are mapped to
suitable ones in the ROOT header les. If you want adaptive code of this type, use the ROOT type declarations.
However, usually you do not need such adaptive code, and you can safely use the standard C type declarations
for your private code, as we did and will do throughout this guide. If you intend to become a ROOT developer,
however, you better stick to the ocial coding rules!

2.7.2. Congure ROOT at start-up

If the le

.rootlogon.C

exists in your home directory, it is executed by ROOT at start-up. Such a le can be

used to set preferred options for each new ROOT session. The ROOT default for displaying graphics looks OK
on the computer screen, but rather ugly on paper. If you want to use ROOT graphs in documents, you should

TStyle object with your preferred

gROOT->SetStyle("MyStyle"); to
in the le rootlogon.C coming with

change some of the default options. This is done most easily by creating a new
settings, as described in the class reference guide, and then use the command
make this new style denition the default one. As an example, have a look
this tutorial.

There is also a possibility to set many ROOT features, in particular those closely related to the operating and
window system, like e.g.

the fonts to be used, where to nd start-up les, or where to store a le containing

the command history, and many others.

10

The le searched for at ROOT start-up is called

.rootrc

and must

2.7. ROOT Beginners' FAQ

reside in the user's home directory; reading and interpeting this le is handled by the ROOT class

TEnv,

see its

documentation if you need such rather advanced features.

2.7.3. ROOT command history

Every command typed at the ROOT prompt is stored in a le

.root_hist

in your home directory.

ROOT

uses this le to allow for navigation in the command history with the up-arrow and down-arrow keys. It is also
convenient to extract successful ROOT commands with the help of a text editor for use in your own macros.

2.7.4. ROOT Global Variables

All global variables in ROOT begin with a small g. Some of them were already implicitly introduced (for example
in session 2.7.2). The most important among them are presented in the following:

gROOT: the gROOT variable is the entry point to the ROOT system.
TROOT

program. The

gRandom:
type

Technically it is an instance of the

class. Using the gROOT pointer one has access to basically every object created in a ROOT based

TROOT

object is essentially a container of several lists pointing to the main

ROOT

objects.

the gRandom variable is a variable that points to a random number generator instance of the

TRandom3.

Such a variable is useful to access in every point of a program the same random number

generator, in order to achieve a good quality of the random sequence.

gStyle:

By default ROOT creates a default style that can be accessed via the

gStyle

pointer. This class

includes functions to set some of the following object attributes.

 Canvas
 Pad
 Histogram axis
 Lines
 Fill areas
 Text
 Markers
 Functions
 Histogram Statistics and Titles
gSystem: An instance of a base class dening a generic interface to the underlying Operating System,
our case

TUnixSystem.

in

At this point you have already learnt quite a bit about some basic features of ROOT.

Please move on to become an expert!

11

CHAPTER 3
ROOT MACROS

You know how other books go on and on about programming fundamentals and nally work up to building a
complete, working program? Let's skip all that. In this part of the guide, we will describe macros executed by
the ROOT C++ interpreter CINT.
An alternative way to access ROOT classes interactively or in a script will be shown in Chapter 8, where we
describe how to use the scritping language

Python.

This is most suitable for smaller analysis projects, as some

overhead of the C++ language can be avoided. It is very easy to convert ROOT macros into python scripts using
the

pyroot interface.

Since ROOT itself is written in C++ let us start with Root macros in C++.

As an additional advantage,

it is relatively easy to turn a ROOT C++ macro into compiled  and hence much faster  code, either as a
pre-compiled library to load into ROOT, or as a stand-alone application, by adding some include statements for
header les or some dressing code to any macro.

3.1. General Remarks on ROOT macros

If you have a number of lines which you were able to execute at the ROOT prompt, they can be turned into a
ROOT macro by giving them a name which corresponds to the le name without extension. The general structure
for a macro stored in le

void MacroName ( )
<

2
3

is

{
...

your lines of CINT code

...

4
5

MacroName.cxx

>

}
The macro is executed by typing

>

root MacroName . cxx

at the system prompt, or it can be loaded into a ROOT session and then be executed by typing
1
2

root
root

[0].
[1]

L MacroName . cxx
MacroName ( ) ;

at the ROOT prompt. Note that more than one macro can be loaded this way, as each macro has a unique name
in the ROOT name space. Because many other macros may have been executed in the same shell before, it is a
good idea to reset all ROOT parameters at the beginning of a macro and dene your preferred graphics options,
e. g. with the code fragment
1
2
3
4
5
6
7
8

/ / r e i n i t i a l i s e ROOT
gROOT >Reset ( ) ;
gROOT >SetStyle ( " Plain " ) ;
gStyle >SetOptStat ( 1 1 1 1 1 1 ) ;
gStyle >SetOptFit ( 1 1 1 1 ) ;
gStyle >SetPalette ( 1 ) ;
gStyle >SetOptTitle ( 0 ) ;

//

re i n i t i a l i z e

//

set

//

print

statistics

//

print

fit

empty

//

set

//

suppress

nicer

ROOT

TStyle

( nicer
on

results
colors
title

on

paper )

plots ,

(0)

for

no

output

plot ,

(0)

for

no

ouput

on
than

default

box

...

13

3. ROOT Macros

Next, you should create a canvas for graphical output, with size, subdivisions and format suitable to your needs,
see documentation of class

1
2
3
4

//

create

TCanvas:

canvas ,

specify

position

and

size

TCanvas c1 ( " c1 " , " <Title >" , 0 , 0 , 4 0 0 , 3 0 0 ) ;

c1 . Divide ( 2 , 2 ) ; / / s e t s u b d i v i s i o n s , c a l l e d
c1 . cd ( 1 ) ; // c h a n g e t o pad 1 o f c a n v a s c 1

in

pixels

pads

These parts of a well-written macro are pretty standard, and you should remember to include pieces of code
like in the examples above to make sure your output always comes out as you had intended.
Below, in section3.4, some more code fragments will be shown, allowing you to use the system compiler to
compile macros for more ecient execution, or turn macros into stand-alone applications linked against the
ROOT libraries.

3.2. A more complete example

Let us now look at a rather complete example of a typical task in data analysis, a macro that constructs a graph
with errors, ts a (linear) model to it and saves it as an image. To run this macro, simply type in the shell:
1

>

root macro1 . cxx

The code is build around the ROOT class

TGraphErrors, which was already introduced previously.

Have a look at

it in the class reference guide, where you will also nd further examples. The macro shown below uses additional
classes,

TLegend

TF1

to dene a function,

TCanvas

to dene size and properties of the window used for our plot, and

to add a nice legend. For the moment, ignore the commented include statements for header les, they

will only become important at the end (section 3.4).

/*

****

//

first ,

3
4
5
6
7
8
9

Builds

# include
# include
# include
# include
# include
# include
# include

graph

include

with

some

errors ,

header

" TCanvas .h"

" TROOT .h"
" TGraphErrors .h"
" TF1 .h"
" TLegend .h"
" TArrow .h"
" TLatex .h"

files

displays
( within

it

and

CINT ,

saves

these

it

as

image .

will

be

ignored )

10
11

void macro1 ( ) {
//

12

The

values

and

the

errors

const int n_points = 1 0 ;

double x_vals [ n_points ]=

13
14

on

the Y a x i s

{1 ,2 ,3 ,4 ,5 ,6 ,7 ,8 ,9 ,10};

15

double y_vals [ n_points ]=

16

{6 ,12 ,14 ,20 ,22 ,24 ,35 ,45 ,44 ,53};

17

double y_errs [ n_points ]=

18

{5 ,5 ,4.7 ,4.5 ,4.2 ,5.1 ,2.9 ,4.1 ,4.8 ,5.43};

19
20
21

//

Instance

of

the

//

Make

plot

//

The

graph

TGraphErrors graph ( n_points , x_vals , y_vals , NULL , y_errs ) ;

graph . SetTitle ( " Measurement XYZ ; lenght [ cm ]; Arb . Units " ) ;

22
23
24
25

the

estetically

better

which

draw

gROOT >SetStyle ( " Plain " ) ;

graph . SetMarkerStyle ( kOpenCircle ) ;
graph . SetMarkerColor ( kBlue ) ;
graph . SetLineColor ( kBlue ) ;

26
27
28
29
30
31

canvas

on

TCanvas * mycanvas

32

33

//

34

Draw

the

graph

graph . DrawClone ( " APE " ) ;

35
36

14

we ' l l

new TCanvas ( ) ;

the

graph

***

*/

3.2. A more complete example

//

Define

38

TF1 f ( " Linear law " , " [0]+ x *[1] "

39

//

Let ' s

40

f . SetLineColor ( kRed ) ; f . SetLineStyle ( 2 ) ;

41

//

Fit

//

Build

//

Draw an

37

linear

make

it

to

the

the

function

funcion

graph

graph . Fit (& f ) ;

f . DrawClone ( " Same " ) ;

42
43

and

,.5 ,10.5) ;

line

draw

nicer
it

44
45

and Draw a

legend

TLegend leg ( . 1 , . 7 , . 3 , . 9 , " Lab . Lesson 1" ) ;

leg . SetFillColor ( 0 ) ;
graph . SetFillColor ( 0 ) ;
leg . AddEntry (& graph , " Exp . Points " ) ;
leg . AddEntry (& f , " Th . Law " ) ;
leg . DrawClone ( " Same " ) ;

46
47
48
49
50
51
52
53

arrow

on

the

canvas

text

to

the

plot

TArrow arrow ( 8 , 8 , 6 . 2 , 2 3 , 0 . 0 2 , " ----|>" ) ;

arrow . SetLineWidth ( 2 ) ;
arrow . DrawClone ( ) ;

54
55
56
57

// Add some

58

TLatex text ( 8 . 2 , 7 . 5 , "# splitline { Maximum }{ Deviation }" ) ;

text . DrawClone ( ) ;

59
60
61

mycanvas >Print ( " graph_with_law . pdf " ) ;

62
63

64
65
66
67

# ifndef __CINT__
int main ( ) {
macro1 ( ) ;
}

68
69

# endif
le:

macro1.cxx

Let's comment it in detail:

Line

11:

the name of the principal function (it plays the role of the main function in compiled programs)

in the macro le. It has to be the same as the le name without extension.

Line

22 23:

instance of the

pointers to the arrays of

and

errors.

TGraphErrors class.
y values, x errors

values,

The constructor takes the number of points and the

(in this case none, represented by the NULL pointer)

The second line denes in one shot the title of the graph and the titles of the two axes,

separated by a ;.

Line

26 29:

the rst line refers to the style of the plot, set as

of the global variable

gSystem

Plain.

This is done through a manipulation

(ROOT global variables begin always with g). The following three lines

are rather intuitive right? To understand better the enumerators for colours and styles see the reference for
the

TColor

Line

32:

and

TMarker

classes.

the canvas object that will host the drawn objects. The memory leak is intentional, to make the

object existing also out of the macro1 scope.

Line

35:

the method

after the scope of

DrawClone

macro1,

draws a clone of the object on the canvas. It

has to be

a clone, to survive

and be displayed on screen after the end of the macro execution.

The string

option APE stands for:

 A imposes the drawing of the Axes.

 P imposes the drawing of the graphs markers.
 E imposes the drawing of the graphs markers errors.

Line

38:

dene a mathematical function. There are several ways to accomplish this, but in this case the

constructor accepts the name of the function, the formula, and the function range.

40: maquillage.
TLine class.
Line

Line

42:

ts the

Try to give a look to the line styles at your disposal visiting the documentation of the

function to the graph, observe that the pointer is passed. It is more interesting to look

at the output on the screen to see the parameters values and other crucial information that we will learn
to read at the end of this guide.

15

3. ROOT Macros

Line

43:

again draws the clone of the object on the canvas. The Same option avoids the cancellation of

the already drawn objects, in our case, the graph.

Line

46 51:

completes the plot with a legend, represented by a

TLegend

instance. The constructor takes

as parameters the lower left and upper right corners coordinates with respect to the total size of the canvas,
assumed to be 1, and the legend header string. You can add to the legend the objects, previously drawn
or not drawn, through the

addEntry

method. Observe how the legend is drawn at the end: looks familiar

now, right?

Line

Line

54 56:
59 61:

denes an arrow with a triangle on the right hand side, a thickness of 2 and draws it.
interpret a Latex string which hast its lower left corner located in the specied coordinate.

The #splitline{}{} construct allows to store multiple lines in the same

Line

62:

TLatex

object.

save the canvas as image. The format is automatically inferred from the le extension (it could

have been eps, gif, . . . ).

Let's give a look to the obtained plot in gure 3.1. Beautiful outcome for such a small bunch of lines, isn't it?

Arb.Units

Measurement XYZ
60

Lab. Lesson 1
Exp. Points

50

Th. Law

40
30
20

Maximum
Deviation

10
0

Figure 3.1.:

Your rst plot with data points.

A version of the same macro in

10
lenght [cm]

Python is available in the le macro1.py; you may want to open it in the

editor and have a look at the dierences right now - please consult the introductory sections of Chapter 8 rst.
This example shows how easy it is to change a ROOT macro from C++ to

Python.

3.3. Summary of Visual eects

3.3.1. Colours and Graph Markers
We have seen that to specify a colour, some identiers like kWhite, kRed or kBlue can be specied for markers,
lines, arrows etc. The complete summary of colours is represented by the ROOT colour wheel, shown in appendix
in gure B.1. To know more about the full story, refer to the online documentation of

TColor.

ROOT provides an analogue of the colour wheel for the graphics markers. Select the most suited symbols for
your plot (see Figure B.1) among dots, triangles, crosses or stars. An alternative set of names for the markers is
summarised in Table B.1.

3.3.2. Arrows and Lines

The macro line 56 shows how to dene an arrow and draw it. The class representing arrows is
inherits from

TLine.

TArrow,

which

The constructors of lines and arrows always contain the coordinates of the endpoints. Arrows

also foresee parameters to specify their shapes (see Figure B.2). Do not underestimate the role of lines and arrows
in your plots.
primitives.

16

Since each plot should contain a message, it is convenient to stress it with additional graphics

3.4. Interpretation and Compilation

3.3.3. Text
Also text plays a fundamental role in making the plots self-explanatory. A possibility to add text in your plot is
provided by the

TLatex

class. The objects of this class are constructed with the coordinates of the bottom-left

corner of the text and a string which contains the text itself. The real twist is that ordinary Latex mathematical
symbols are automatically interpreted, you just need to replace the \ by a # (see Figure B.3).

3.4. Interpretation and Compilation

As you observed, up to now we heavily exploited the capabilities of ROOT for interpreting our code, more than
compiling and then executing.

This is sucient for a wide range of applications, but you might have already

asked yourself how can this code be compiled?. There are two answers.

3.4.1. Compile a Macro with ACLiC

ACLiC will create for you a compiled dynamic library for your macro, without any eort from your side, except
the insertion of the appropriate header les in lines 39. In this example, they are already included. This does
not harm, as they are not loaded by CINT. To generate an object libary from the macro code, from inside the
interpreter type (please note the +):

root

[1]

L macro1 . cxx+

Once this operation is accomplished, the macro symbols will be available in memory and you will be able to
execute it simply by calling from inside the interpreter:

root

[2]

macro1 ( )

3.4.2. Compile a Macro with g++

In this case, you have to include the appropriate headers in the code and then exploit the
the automatic settings of all the compiler ags.

root-cong

root-cong

tool for

is a script that comes with ROOT; it prints all ags

and libraries needed to compile code and link it with the ROOT libraries. In order to make the code executable
stand-alone, an entry point for the operating system is needed, in C++ this is the procedure

int main();.

The

easiest way to turn a ROOT macro code into a stand-alone application is to add the following dressing code at
the end of the macro le. This denes the procedure main, the only purpose of which is to call your macro:
1
2
3
4
5
6

# ifndef __CINT__
int main ( ) {
ExampleMacro ( ) ;
return 0 ;
}

# endif
Within ROOT, the symbol

__CINT__

#ifndef __CINT__ and #endif is not

g++, this symbol is not dened, and the code is
called ExampleMacro.C, simply type

is dened, and the code enclosed by

executed; on the contrary, when running the system compiler

compiled. To create a stand-alone program from a macro
1

>

g++ o ExampleMacro . exe ExampleMacro . C

rootconfig cflags libs `

and execute it by typing

1

>

./

ExampleMacro . exe

This procedure will, however, not give access to the ROOT graphics, as neither control of mouse or keyboard
events nor access to the graphics windows of ROOT is available. If you want your stand-alone application have
display graphics output and respond to mouse and keyboard, a slightly more complex piece of code can be used.
In the example below, a macro

ExampleMacro_GUI

is executed by the ROOT class TApplication. As a further

feature, this code example oers access to parameters eventually passed to the program when started from the
command line. Here is the code fragment:
1
2

# ifndef __CINT__
void StandaloneApplication ( int argc , char * * argv )

//

/ / ==>> h e r e

evaluate

the

t h e ROOT macro

ExampleMacro_GUI ( ) ;

5
6

eventually ,

application
is

parameters

argc ,

argv

called

17

3. ROOT Macros

//

7
8
9
10
11
12
13
14
15

This

is

the

standard

" main "

of

C++ s t a r t i n g

a ROOT a p p l i c a t i o n

int main ( int argc , char * * argv ) {

gROOT >Reset ( ) ;
TApplication app ( " Root Application " , &argc , argv ) ;
StandaloneApplication ( app . Argc ( ) , app . Argv ( ) ) ;
app . Run ( ) ;
return 0 ;
}

# endif
Compile the code with

>

g++ o ExampleMacro_GUI . exe ExampleMacro_GUI

and execute the program with

1

>

18

./

ExampleMacro_GUI . exe

rootconfig cflags libs `

CHAPTER 4
GRAPHS

In this Chapter we will learn how to exploit some of the functionalities that ROOT provides to display data based
on the class

TGraphErrors,

which you already got to know previously.

4.1. Read Graph Points from File

The fastest way in which you can ll a graph with experimental data is to use the constructor which reads data
points and their errors from a le in ASCII (i. e. standard text) format:

TGraphErrors ( const char * filename , const char

Option_t * option="" ) ;

format="% lg % lg % lg % lg " , -

The format string can be:

"\%lg \%lg"

"\%lg \%lg \%lg"

"\%lg \%lg \%lg \%lg"

read only 2 rst columns into X,Y

read only 3 rst columns into X,Y and EY
read only 4 rst columns into X,Y,EX,EY

This approach has a the nice feature of allowing the user to reuse the macro for many dierent data sets. Here
is an example of an input le. The nice graphic result shown is produced by the macro below, which reads two
such input les and uses dierent options to display the data points.

Measurement XYZ and Expectation

2

Measurement
Experiment

of
2

Friday

Physics

26

March

Arb.Units

#
#

Lab

3
4

12

14

4.7

20

4.5

22

4.2

24

5.1

10

35

2.9

11

45

4.1

12

44

4.8

13

10

53

5.43

/*

Reads

the

int macro2 ( )

4
5
6

Expected Points

60

Measured Points

50
40
30
20
10
le:

Lab. Lesson 2

70

points

from

macro2_input.txt

file

and

produces

simple

graph .

10
lenght [cm]

*/

gROOT >SetStyle ( " Plain " ) ;

TCanvas * c=new TCanvas ( ) ;
c>SetGrid ( ) ;

19

4. Graphs

TGraphErrors graph_expected ( " ./ macro2_input_expected . txt " , "% lg % lg % lg " ) ;

graph_expected . SetTitle ( " Measurement XYZ and Expectation ; lenght [ cm ]; Arb . Units " ) ;
graph_expected . SetFillColor ( kYellow ) ;
graph_expected . DrawClone ( " E3AL " ) ; // E3 d r a w s t h e band

8
9

10
11
12

TGraphErrors graph ( " ./ macro2_input . txt " , "% lg % lg % lg " ) ;

graph . SetMarkerStyle ( kCircle ) ;
graph . SetFillColor ( 0 ) ;
graph . DrawClone ( " PESame " ) ;

13
14
15
16
17

//

18

Draw

the

Legend

TLegend leg ( . 1 , . 7 , . 3 , . 9 , " Lab . Lesson 2" ) ;

leg . SetFillColor ( 0 ) ;
leg . AddEntry (& graph_expected , " Expected Points " ) ;
leg . AddEntry (& graph , " Measured Points " ) ;
leg . DrawClone ( " Same " ) ;

19
20
21
22
23
24

c>Print ( " graph_with_band . pdf " ) ;

return 0 ;

25
26
27

}
le:
Beyond looking at the plot, you can check the actual contents of the graph with the

macro2.cxx

TGraph::Print()

method

at any time, obtaining a printout of the coordinates of data points on screen. The macro also shows us how to
print a coloured band around a graph instead of error bars, quite useful for example to represent the errors of a
theoretical prediction.

4.2. Polar Graphs

With ROOT you can prot from rather advanced plotting routines, like the ones implemented in the

TPolarGraph,

a class to draw graphs in polar coordinates. It is very easy to use, as you see in the example macro and the resulting
plot 4.1:

/*

*/

3
4
5
6
7
8
9
10
11

polar

graph

in

square

Canvas

void macro3 ( ) {
double rmin =0;
double rmax=TMath : : Pi ( ) * 6 ;
const int npoints = 3 0 0 ;
Double_t r [ npoints ] ;
Double_t theta [ npoints ] ;
for ( Int_t ipt = 0 ; ipt < npoints ; ipt ++) {
r [ ipt ] = ipt * ( rmaxrmin ) / ( npoints 1 . )+rmin ;
theta [ ipt ] = TMath : : Sin ( r [ ipt ] ) ;
}

12

TCanvas * c = new TCanvas ( " myCanvas " , " myCanvas " , 6 0 0 , 6 0 0 ) ;

TGraphPolar grP1 ( npoints , r , theta ) ;
grP1 . SetTitle ( "A Fan " ) ;
grP1 . SetLineWidth ( 3 ) ;
grP1 . SetLineColor ( 2 ) ;
grP1 . DrawClone ( " AOL " ) ;

13
14
15
16
17
18
19

Builds

}
le:
A new element was added on line 4, the size of the canvas:
in specic canvas sizes.
Some

20

macro3.cxx

it is sometimes optically better to show plots

Python variants of this macro are shown and discussed in Chapter 8.

4.3. 2D Graphs

A Fan

3
4

0
-1

-0.5

5
4

0.5

7
4
3
2

Figure 4.1.:

The graph of a fan obtained with ROOT.

4.3. 2D Graphs
On some occasions it might be useful to plot some quantities versus two variables, therefore creating a bidimensional graph. Of course ROOT can help you in this task, with the

TGraph2DErrors

class. The following

macro produces a bi-dimensional graph representing a hypothetical measurement, ts a bi-dimensional function
to it and draws it together with its x and y projections. Some points of the code will be explained in detail. This
time, the graph is populated with data points using random numbers, introducing a new and very important
ingredient, the ROOT

1
2
3
4

/*

Create ,

TRandom3

Draw and

fit

random number generator using the Mersenne Twister algorithm [6].

TGraph2DErrors

void macro4 ( ) {
gStyle >SetPalette ( 1 ) ;
gROOT >SetStyle ( " Plain " ) ;

*/

const double e
const int nd =

6
7

0.3;

500;

TRandom3 my_random_generator ;
TF2 * f2 = new TF2 ( " f2 " , " 1000*(([0]* sin (x)/x) *([1]* sin (y)/y)) +200 " , 6 ,6 , 6 ,6) ;
f2>SetParameters ( 1 , 1 ) ;
TGraph2DErrors * dte = new TGraph2DErrors ( nd ) ;

9
10

11
12
13

//

14
15
16
17
18
19
20
21
22
23

26
27
28

the

2D g r a p h

double rnd , x , y , z , ex , ey , ez ;
for ( Int_t i =0; i<nd ; i++) {
f2>GetRandom2 ( x , y ) ;
rnd = my_random_generator . Uniform ( e , e ) ;
z = f2>Eval ( x , y ) * (1+ rnd ) ;
dte>SetPoint ( i , x , y , z ) ;
ex = 0 . 0 5 * my_random_generator . Uniform ( ) ;
ey = 0 . 0 5 * my_random_generator . Uniform ( ) ;
ez = TMath : : Abs ( z * rnd ) ;
dte>SetPointError ( i , ex , ey , ez ) ;

// A random

number

in

[e , e ]

24
25

Fill

//

Fit

function

to

generated

data

f2>SetParameters ( 0 . 7 , 1 . 5 ) ; // s e t
f2>SetTitle ( " Fitted 2D function " ) ;
dte>Fit ( f2 ) ;

initial

values

for

fit

21

4. Graphs

29

//

Plot

the

result

//

Make

the

TCanvas * c1 = new TCanvas ( ) ;

f2>Draw ( " Surf1 " ) ;
dte>Draw ( " P0 Same " ) ;

30
31
32
33

35
36
37
38
39
40

and

projections

TCanvas * c_p= new TCanvas ( " ProjCan " , " The Projections " , 1 0 0 0 , 4 0 0 ) ;
c_p>Divide ( 2 , 1 ) ;
c_p>cd ( 1 ) ;
dte>Project ( "x" )>Draw ( ) ;
c_p>cd ( 2 ) ;
dte>Project ( "y" )>Draw ( ) ;

34

}
le:

Line

3:

macro4.cxx

This sets the palette colour code to a much nicer one than the default. Comment this line to give

it a try.

Line

4:

sets a style type without ll color and shadows for pads.

Looks much nicer on paper than the

default setting.

Line

9:

The instance of the random generator. You can then draw out of this instance random numbers

distributed according to dierent probability density functions, like the Uniform one at lines 25,26. See the
on-line documentation to appreciate the full power of this ROOT feature.

Line

10:

You are already familiar with the

TF1

class. This is its two-dimensional correspondent. At line 21

two random numbers distributed according to the

TF2::GetRandom2(double& a, double&b).

TF2

formula are drawn with the method

Line 2628: Fitting a 2-dimensional function just works like in the one-dimensional case, i.e. initialisation

Line

of parameters and calling of the

31:

The

Surf1

Fit()

option draws the

method.

TF2

objects (but also bi-dimensional histograms) as coloured surfaces

with a wire-frame on three-dimensional canvases.

Line

3439:

Here you learn how to create a canvas, partition it in two sub-pads and access them. It is very

handy to show multiple plots in the same window or image.

22

CHAPTER 5
HISTOGRAMS

Histograms play a fundamental role in any type of Physics analysis, not only displaying measurements but being
a powerful form of data reduction. ROOT presents many classes that represent histograms, all inheriting from the

TH1 class.

We will focus in this chapter on uni- and bi- dimensional histograms whose bin-contents are represented

by oating point numbers

, the

TH1F

and

TH2F

classes respectively.

5.1. Your First Histogram

Let's suppose that you want to measure the counts of a Geiger detector put in proximity of a radioactive source
in a given time interval. This would give you an idea of the activity of your source. The count distribution in
this case is a Poisson distribution. Let's see how operatively you can ll and draw a histogram in the following
example macro.

/* C r e a t e ,

Fill

counts

of

and

scaler

draw

an

linked

Histogram
to

which

Geiger

reproduces

the

c o u n t e r . */

3
4
5
6

void macro5 ( ) {
TH1F * cnt_r_h=new TH1F ( " count_rate " ,
" Count Rate ; N_ { Counts };# occurencies " ,

100 ,

0.5 ,

//

15.5) ;

Number

//
//

of

Bins

Lower X Boundary
Upper X Boundary

10

12

const float mean_count = 3 . 6 ;

TRandom3 rndgen ;

13

//

11

for

14

simulate
( int

the

measurements

imeas =0; imeas < 4 0 0 ; imeas ++)

cnt_r_h >Fill ( rndgen . Poisson ( mean_count ) ) ;

15
16

gROOT >SetStyle ( " Plain " ) ;

TCanvas * c= new TCanvas ( ) ;
cnt_r_h >Draw ( ) ;

17
18
19
20

TCanvas * c_norm= new TCanvas ( ) ;

cnt_r_h >DrawNormalized ( ) ;

21
22
23

//

24

Print

cout

25

<<
<<

26
27

<<

28
29

<<

30

<<

31
32

summary

" Moments of Distribution :\ n"

" - Mean = " << cnt_r_h >GetMean ( ) << " +- "
<< cnt_r_h >GetMeanError ( ) << "\n"
" - RMS = " << cnt_r_h >GetRMS ( ) << " +- "
<< cnt_r_h >GetRMSError ( ) << "\n"
" - Skewness = " << cnt_r_h >GetSkewness ( ) << "\n"
" - Kurtosis = " << cnt_r_h >GetKurtosis ( ) << "\n" ;

}
le:

1 To

macro5.cxx

optimise the memory usage you might go for one byte (TH1C), short (TH1S), integer (TH1I) or double-precision

(TH1D) bin-content.

23

5. Histograms

count_rate
Entries
400
Mean
3.562
RMS
1.792

# occurencies

Count Rate
90
80
70
60
50
40
30
20
10
0

Figure 5.1.:

10

12

14
NCounts

The result of a counting (pseudo) experiment.

Which gives you the following plot 5.1: Using histograms is rather simple. The main dierences with respect to
graphs that emerge from the example are:

line 5: The histograms have a name and a title right from the start, no predened number of entries but a
number of bins and a lower-upper range.

line 15: An entry is stored in the histogram through the

line 19 and 22:

TH1F::Fill

method.

The histogram can be drawn also normalised, ROOT automatically takes cares of the

necessary rescaling.

line 25 to 31: This small snippet shows how easy it is to access the moments and associated errors of a
histogram.

5.2. Add and Divide Histograms

Quite a large number of operations can be carried out with histograms. The most useful are addition and division.
In the following macro we will learn how to manage these procedures within ROOT.

/* D i v i d e

and

add

1D H i s t o g r a m s * /

2
3
4
5

void format_h ( TH1F * h , int linecolor ) {

h>SetLineWidth ( 3 ) ;
h>SetLineColor ( linecolor ) ;
}

6
7
8
9

void macro6 ( ) {
gROOT >SetStyle ( " Plain " ) ;

10

TH1F *
TH1F *
TH1F *
TH1F *

11
12
13
14

sig_h=new TH1F ( " sig_h " , " Signal Histo " , 5 0 , 0 , 1 0 ) ;

gaus_h1=new TH1F ( " gaus_h1 " , " Gauss Histo 1" , 3 0 , 0 , 1 0 ) ;
gaus_h2=new TH1F ( " gaus_h2 " , " Gauss Histo 2" , 3 0 , 0 , 1 0 ) ;
bkg_h=new TH1F ( " exp_h " , " Exponential Histo " , 5 0 , 0 , 1 0 ) ;

15
16

//

simulate

//

Format

the

measurements

TRandom3 rndgen ;
for ( int imeas =0; imeas < 4 0 0 0 ; imeas ++){
bkg_h >Fill ( rndgen . Exp ( 4 ) ) ;
if ( imeas %4==0) gaus_h1 >Fill ( rndgen . Gaus ( 5 , 2 ) ) ;
if ( imeas %4==0) gaus_h2 >Fill ( rndgen . Gaus ( 5 , 2 ) ) ;
if ( imeas %10==0) sig_h >Fill ( rndgen . Gaus ( 5 , . 5 ) ) ; }

17
18
19
20
21
22
23
24

Histograms

TH1F * histos [ 4 ] = { sig_h , bkg_h , gaus_h1 , gaus_h2 } ;

for ( int i =0; i <4;++i ) {
histos [ i ]> Sumw2 ( ) ; // * Very * I m p o r t a n t

25
26
27

24

Gaus Histo 1 and Gaus Histo 2

5.2. Add and Divide Histograms

220
200
180

0.07

0.06

0.05

0.04

0.03

160

0.02

140

0.01

Gaus Histo 1 / Gaus Histo 2

120
100
80
60
40

2.4 0

10

10
X axis

2.2
2
1.8
1.6
1.4
1.2
1
0.8

20

0.6
0.4

Figure 5.2.:

10

0.2
0
0

The sum of two histograms and the ratio.

28

format_h ( histos [ i ] , i +1) ;

29

30

// Sum

31

TH1F * sum_h= new TH1F ( * bkg_h ) ;

sum_h >Add ( sig_h , 1 . ) ;
sum_h >SetTitle ( " Exponential + Gaussian " ) ;
format_h ( sum_h , kBlue ) ;

32
33
34
35
36

TCanvas * c_sum= new TCanvas ( ) ;

sum_h >Draw ( " hist " ) ;
bkg_h >Draw ( " SameHist " ) ;
sig_h >Draw ( " SameHist " ) ;

37
38
39
40
41
42

//

Divide

//

Graphical

TH1F * dividend=new TH1F ( * gaus_h1 ) ;

dividend >Divide ( gaus_h2 ) ;

43
44
45
46

Maquillage

dividend >SetTitle ( ";X axis ; Gaus Histo 1 / Gaus Histo 2" ) ;

format_h ( dividend , kOrange ) ;
gaus_h1 >SetTitle ( " ;; Gaus Histo 1 and Gaus Histo 2" ) ;
gStyle >SetOptStat ( 0 ) ;
gStyle >SetOptTitle ( 0 ) ;

47
48
49
50
51
52

TCanvas * c_divide= new TCanvas ( ) ;

c_divide >Divide ( 1 , 2 , 0 , 0 ) ;
c_divide >cd ( 1 ) ;
c_divide >GetPad ( 1 )>SetRightMargin ( . 0 1 ) ;
gaus_h1 >DrawNormalized ( " Hist " ) ;
gaus_h2 >DrawNormalized ( " HistSame " ) ;
c_divide >cd ( 2 ) ;
dividend >GetYaxis ( )>SetRangeUser ( 0 , 2 . 4 9 )
c_divide >GetPad ( 2 )>SetGridy ( ) ;
c_divide >GetPad ( 2 )>SetRightMargin ( . 0 1 ) ;
dividend >Draw ( ) ;

53
54
55
56
57
58
59
60
61
62
63
64

}
le:

macro6.cxx

The plots that you will obtain are shown in 5.2 Some lines now need a bit of clarication:

line 3: CINT, as we know, is also able to interpret more than one function per le. In this case the function
simply sets up some parameters to conveniently set the line of histograms.

line 20 to 22: Some contracted C++ syntax for conditional statements is used to ll the histograms with
dierent numbers of entries inside the loop.

line 27: This is a crucial step for the sum and ratio of histograms to handle errors properly. The method

TH1::Sumw2

causes the squares of weights to be stored inside the histogram (equivalent to the number of

25

5. Histograms

entries per bin if weights of 1 are used). This information is needed to correctly calculate the errors of each
bin entry when the methods

TH1::Add

line 33: The sum of two histograms.

and

TH1::Divide

are applied.

A weight can be assigned to the added histogram, for example to

comfortably switch to subtraction.

line 44: The division of two histograms is rather straightforward.

line 53 to 63: When you draw two quantities and their ratios, it is much better if all the information is
condensed in one single plot. These lines provide a skeleton to perform this operation.

5.3. Two-dimensional Histograms

Two-dimensional histograms are a very useful tool, for example to inspect correlations between variables. You
can exploit the bi-dimensional histogram classes provided by ROOT in a very simple way. Let's see how in the
following macro:

/*

together

Draw a

Bidimensional
with

its

Histogram

profiles

and

in

many ways

p r o j e c t i o n s */

3
4
5
6
7
8

void macro7 ( ) {
gROOT >SetStyle ( " Plain " ) ;
gStyle >SetPalette ( 1 ) ;
gStyle >SetOptStat ( 0 ) ;
gStyle >SetOptTitle ( 0 ) ;

TH2F bidi_h ( " bidi_h " ,

"2D Histo ; Guassian Vals ; Exp . Vals " ,
30 , 5 ,5 ,
// X a x i s

10
11
12

30 ,0 ,10) ;

13

// Y a x i s

14

TRandom3 rndgen ;
for ( int i =0; i < 5 0 0 0 0 0 ; i++)
bidi_h . Fill ( rndgen . Gaus ( 0 , 2 ) ,
10 rndgen . Exp ( 4 ) ) ;

15
16
17
18
19

TCanvas * c=new TCanvas ( " Canvas " , " Canvas " , 8 0 0 , 8 0 0 ) ;

c>Divide ( 2 , 2 ) ;
c>cd ( 1 ) ; bidi_h . DrawClone ( " Contz " ) ;
c>cd ( 2 ) ; bidi_h . DrawClone ( " Colz " ) ;
c>cd ( 3 ) ; bidi_h . DrawClone ( " lego2 " ) ;
c>cd ( 4 ) ; bidi_h . DrawClone ( " surf3 " ) ;

20
21
22
23
24
25
26

//

27

29
30
31
32
33
34

Profiles

and

Projections

TCanvas * c2=new TCanvas ( " Canvas2 " , " Canvas2 " , 8 0 0 , 8 0 0 ) ;

c2>Divide ( 2 , 2 ) ;
c2>cd ( 1 ) ; bidi_h . ProjectionX ( )>DrawClone ( ) ;
c2>cd ( 2 ) ; bidi_h . ProjectionY ( )>DrawClone ( ) ;
c2>cd ( 3 ) ; bidi_h . ProfileX ( )>DrawClone ( ) ;
c2>cd ( 4 ) ; bidi_h . ProfileY ( )>DrawClone ( ) ;

28

}
le:

macro macro7.cxx

Two kinds of plots are provided by the code, the rst one containing three-dimensional representations (Figure 5.3) and the second one projections and proles (5.4) of the bi-dimensional histogram.

When a projection

is performed along the x (y) direction, for every bin along the x (y) axis, all bin contents along the y (x) axis are
summed up (upper the plots of gure 5.4). When a prole is performed along the x (y) direction, for every bin
along the x (y) axis, the average of all the bin contents along the y (x) is calculated together with their RMS and
displayed as a symbol with error bar (lower two plots of gure). 5.4).
Correlations between the variables are quantied by the methods
and

26

Double_t GetCorrelationFactor().

Double_T GetCovariance()

10
2500

9
8

Exp. Vals

Exp. Vals

5.3. Two-dimensional Histograms

10

2000

2500

2000

6
1500

1500
5

1000

1000
3

500

500

0
-5

-4

-3

-2

-1

0
-5

2
3
4
5
Guassian Vals

3000

3000

2500

2500

2000

2000

1500

1500

1000

1000

500

500

0
10E
x9p

.8V
a7ls
6

Figure 5.3.:

0
-2 -1
0 -5 -4 -3

5
3 4Vals
1 ua2ssian
G

-4

0
10
Ex9
p. 8
Va7
ls6

-3

-2

-1

2
3
4
5
Guassian Vals

0
-2 -1
0 -5 -4 -3

5
3 4
Vals
1 ua2ssian
G

Dierent ways of representing bi-dimensional histograms.

35000

40000

30000

35000

25000

30000

20000

25000
20000

15000

15000
10000
10000
5000
5000
-5

-4

-3

-2

-1

3
4
5
Guassian Vals

9
10
Exp. Vals

9
10
Exp. Vals

6.96
6.94

0.06

6.92

0.04

6.9
0.02
6.88
6.86

6.84

-0.02

6.82
-0.04
6.8
-5

-4

Figure 5.4.:

-3

-2

-1

3
4
5
Guassian Vals

-0.06
0

The projections and proles of bi-dimensional histograms.

27

CHAPTER 6
FILE I/O

6.1. Storing ROOT Objects

TObject (basically
TFile class. One says

ROOT oers the possibility to write the instances of all the classes inheriting from the class
all classes in ROOT) on disk, into what is referred to as

ROOT-le,

a le created by the

that the object is made persistent by storing it on disk. When reading the le back, the object can be restored
to memory.
We can explore this functionality with histograms and two simple macros.

void write_to_file ( ) {

2
3

//

Istance

of

our

histogram

//

Let ' s

fill

it

//

Let ' s

open

//

Write

the

histogram

//

Close

the

file

TH1F h ( " my_histogram " , " My Title ;X ;# of entries " , 1 0 0 , 5 , 5 ) ;

4
5
6

randomly

h . FillRandom ( " gaus " ) ;

7
8
9

TFile

TFile out_file ( " my_rootfile . root " , " RECREATE " ) ;

10
11
12

h . Write ( ) ;

13

in

the

file

14
15

out_file . Close ( ) ;

16
17

}
le:
The

RECREATE

write_to_file.cxx

option forces ROOT to create a new le even if a le with the same name exists on disk.

Now, you may use the CINT command line to access information in the le and draw the previously written
histogram:

1
2
3
4
5
6
7
8

root my_rootfile . root

root [ 0 ]
Attaching file my_rootfile . root as _file0
root [ 1 ] _file0 . ls ( )
TFile * *
my_rootfile . root
TFile *
my_rootfile . root
KEY : TH1F
my_histogram ; 1 My Title
root [ 2 ] my_histogram . Draw ( )

>>>

...

Alternatively, you can use a simple macro to carry out the job:

29

6. File I/O

void read_from_file ( ) {

2
3

//

Let ' s

//

Get

//

Draw

open

the

TFile

TFile * in_file= new TFile ( " my_rootfile . root " ) ;

4
5
6

TH1F * h

the

Histogram

=( TH1F * )

out

in_file >GetObjectChecked ( " my_histogram " , " TH1F " ) ;

8
9

it

h>DrawClone ( ) ;

10
11
12

}
le:

read_from_file.cxx

Please note that the order of opening les for write access and creating objects determines whether the objects are stored or not. You can avoid this behaviour by using the

Write()

function as shown in the previous

example.
Although you could access an object within a le also with the
advisable to use

GetObjectChecked.

Get

function and a dynamic type cast, it is

6.2. N-tuples in ROOT

6.2.1. Storing simple N-tuples
Up to now we have seen how to manipulate input read from ASCII les. ROOT oers the possibility to do much
better than that, with its own n-tuple classes. Among the many advantages provided by these classes one could
cite

Optimised disk I/O.

Possibility to store many n-tuple rows (Millions).

Write the n-tuples in ROOT les.

Interactive inspection with

Store not only numbers, but also

TBrowser.

objects

In this section we will discuss briey the

ROOT

TNtuple

in the columns.

TNtuple

class, which is a simplied version of the

TTree

class.

commenting a minimal example

/*
an n t u p l e

Fill

conductivity

*/

of

and
a

write

it

material

to

in

file

simulating

different

measurement

conditions

of

pressure

of
and

temperature .

5
6

void write_ntuple_to_file ( ) {

7
8

//

Initialise

//

Fill

it

add

some

random

the

TNtuple

TNtuple cond_data ( " cond_data " ,

" Example N - Tuple " ,
" Potential : Current : Temperature : Pressure " ) ;

9
10
11
12
13

15
16
17
18
19
20

randomly

to

fake

the

acquired

data

float pot , cur , temp , pres ;

for ( int i =0; i <10000;++ i ) {
pot=gRandom >Uniform ( 0 . , 1 0 . ) ;
// g e t v o l t a g e
temp=gRandom >Uniform ( 2 5 0 . , 3 5 0 . ) ; / / g e t t e m p e r a t u r e
pres=gRandom >Uniform ( 0 . 5 , 1 . 5 ) ;
// g e t p r e s s u r e
cur=pot / ( 1 0 . + 0 . 0 5 * ( temp 3 0 0 . ) 0 . 2 * ( pres 1 . ) ) ; // c a l c u l a t e

14

//

21
22
23
24

30

object can store rows of oat entries. Let's tackle the problem according to the usual strategy

smearing

( measurement

pot *= gRandom >Gaus ( 1 . , 0 . 0 1 ) ;

temp+=gRandom >Gaus ( 0 . , 0 . 3 ) ;
pres *= gRandom >Gaus ( 1 . , 0 . 0 2 )
cur *= gRandom >Gaus ( 1 . , 0 . 0 1 ) ;

// 1%
//
;

current

errors )
0.3

error

on

absolute

voltage
error

on

// 1%

error

on

pressure

// 1%

error

on

current

temperature

6.2. N-tuples in ROOT

25

//

write

to

ntuple

26

cond_data . Fill ( pot , cur , temp , pres ) ;

27

28

//

29

Open a

file ,

save

the

ntuple

and

close

the

file

32

TFile ofile ( " conductivity_experiment . root " , " RECREATE " ) ;

cond_data . Write ( ) ;
ofile . Close ( ) ;

33

30
31

le:

write_ntuple_to_file.cxx

This data written to this example n-tuple represents, in the statistical sense, three independent variables (Potential or Voltage, Pressure and Temperature), and one variable (Current) which depends on the the others according
to very simple laws, and an additional Gaussian smearing.

This set of variables mimics a measurement of an

electrical resistance while varying pressure and temperature.

Imagine your task now consists in nding the relations among the variables  of course without knowing the
code used to generate them. You will see that the possibilities of the

NTuple

class enable you to perform this

analysis task. Open the ROOT le (cond_data.root) written by the macro above in an interactive section and
use a

TBrowser

to interactively inspect it:

root [ 0 ] new TBrowser ( )

You nd the columns of your n-tuple written as

leafs.

Simply clicking on them you can obtain histograms of

the variables!
Next, try the following commands at the shell prompt and in the interactive ROOT shell, respectively:
1
2
3

root conductivity_experiment . root

Attaching file conductivity_experiment . root as _file0
root [ 0 ] cond_data . Draw ( " Current : Potential " )

>

...

You just produced a correlation plot with one single line of code!
Try to extend the syntax typing for example
1

root

cond_data . Draw ( " Current : Potential " , " Temperature <270 " )

[1]

What do you obtain?

Now try
1

root

cond_data . Draw ( " Current / Potential : Temperature " )

[2]

It should have become clear from these examples how to navigate in such a multi-dimensional space of variables
and uncover relations between variables using n-tuples.

6.2.2. Reading N-tuples

For completeness, you nd here a small macro to read the data back from a ROOT n-tuple

/*

Read

*/

the

previously

p r o d u c e d NT u p l e

and

print

on

screen

its

content

4
5

void read_ntuple_from_file ( ) {

6
7
8
9

//

Open a

file ,

save

the

ntuple

and

close

the

file

TFile in_file ( " conductivity_experiment . root " ) ;

TNtuple * my_tuple = ( TNtuple * ) in_file . GetObjectChecked ( " cond_data " , " TNtuple " );

10
11
12

float pot , cur , temp , pres ;

float * row_content ;

13
14
15
16
17
18

cout << " Potential \ tCurrent \ tTemperature \ tPressure \n" ;

for ( int irow =0; irow<my_tuple >GetEntries ( ) ;++ irow ) {
my_tuple >GetEntry ( irow ) ;
row_content = my_tuple >GetArgs ( ) ;
pot = row_content [ 0 ] ;

31

6. File I/O

22

cur = row_content [ 1 ] ;
temp = row_content [ 2 ] ;
pres = row_content [ 3 ] ;
cout << pot << "\t" << cur

23

19
20
21

<<

"\t"

<<

temp

<<

"\t"

pres

<<

<<

endl ;

24

25

le:

read_ntuple_from_file.cxx

The macro shows the easiest way of accessing the content of a n-tuple: after loading the n-tuple, its branches are
assigned to variables and

GetEntry(long)

automatically lls them with the content for a specic row. By doing

so, the logic for reading the n-tuple and the code to process it can be split and the source code remains clear.

6.2.3. Storing Arbitrary N-tuples

It is also possible to write n-tuples of arbitrary type by using ROOT's

TNtuple::Fill()

as

are booked directly. The

TBranch class.

This is especially important

accepts only oats. The following macro creates the ame n-tuple as before but the branches

Fill()

function then lls the current values of the connected variables to the tree.

/*
an n t u p l e

Fill

conductivity

using

*/

of

and
a

write

it

material

in

to

file

simulating

different

conditions

measurement
of

of

pressure

and

temperature .

branches

6
7

void write_ntuple_to_file_advanced ( std : : string outputFileName

conductivity_experiment . root " , unsigned int numDataPoints

//

Initialise

the

//

define

variables

" -

= 10000) {

TNtuple

TTree cond_data ( " cond_data " , " Example N - Tuple " ) ;

9
10
11

the

and

book

them

for

the

ntuple

float pot , cur , temp , pres ;

cond_data . Branch ( " Potential " , &pot , " Potential /F" ) ;
cond_data . Branch ( " Current " , &cur , " Current /F" ) ;
cond_data . Branch ( " Temperature " , &temp , " Temperature /F" ) ;
cond_data . Branch ( " Pressure " , &pres , " Pressure /F" ) ;

12
13
14
15
16
17

for

18

( int

//

19

i =0; i<numDataPoints ;++ i ) {

Fill

it

randomly

to

fake

the

acquired

data

pot=gRandom >Uniform ( 0 . , 1 0 . ) * gRandom >Gaus ( 1 . , 0 . 0 1 ) ;

temp=gRandom >Uniform ( 2 5 0 . , 3 5 0 . )+gRandom >Gaus ( 0 . , 0 . 3 ) ;
pres=gRandom >Uniform ( 0 . 5 , 1 . 5 ) * gRandom >Gaus ( 1 . , 0 . 0 2 ) ;
cur=pot / ( 1 0 . + 0 . 0 5 * ( temp 3 0 0 . ) 0 . 2 * ( pres 1 . ) ) * gRandom >Gaus ( 1 .

20
21
22
23

,0.01) ;

24

//

25

write

to

ntuple

26

cond_data . Fill ( ) ;

27

28

//

29

31
32
33

Open a

file ,

save

the

ntuple

and

close

the

file

TFile ofile ( outputFileName . c_str ( ) , " RECREATE " ) ;

cond_data . Write ( ) ;
ofile . Close ( ) ;

30

}
le:

The

Branch()

write_ntuple_to_file_advanced.cxx

function requires a pointer to a variable and a denition of the variable type. Table 6.1 lists

some of the possible values. Please note that ROOT is not checking the input and mistakes are likely to result in
serious problems. This holds especially if values are read as another type than they have been written, e.g. when
storing a variable as oat and reading it as double.

32

6.2. N-tuples in ROOT

Table 6.1.: List of variable types that can be used to dene the type of a branch in ROOT.

type

size

signed integer
unsigned integer
oating point

C++

identier

32 bit

int

64 bit

long

32 bit

unsigned int

64 bit

unsigned long

32 bit

oat

64 bit

double

bool

boolean

6.2.4. Processing N-tuples Spanning over Several Files

Usually n-tuples or trees span over many les and it would be dicult to add them manually. ROOT thus kindly

TChain. Its usage is shown in the following macro which is very similar to
TChain takes the name of the TTree (or TNuple) as an argument. The
Add(fileName), where one can also use wild-cards as shown in the example.

provides a helper class in the form of

the previous example. The constructor of a

les are added with the function

/*
Read

several

previously

p r o d u c e d NT u p l e s

and

print

on

screen

its

content

3
4

you

can

for

easily

in

create

some

do

5;

files

with

the

l x b q

root

conductivity_experiment_$ { i } . root \" ,

6

following

statement :

" w r i t e _ n t u p l e _ t o _ f i l e . c x x ( \ " -

100) " ;

done

*/

7
8

void read_ntuple_with_chain ( ) {
//

initiate

TChain

with

the

name

of

the

TTree

TChain in_chain ( " cond_data " ) ;

in_chain . Add ( " conductivity_experiment *. root " ) ;

10
11

to
//

be
add

processed
files ,

wildcards

work

12

//

13

define

variables

and

assign

them

to

the

corresponding

branches

float pot , cur , temp , pres ;

my_tuple >SetBranchAddress ( " Potential " , &pot ) ;
my_tuple >SetBranchAddress ( " Current " , &cur ) ;
my_tuple >SetBranchAddress ( " Temperature " , &temp ) ;
my_tuple >SetBranchAddress ( " Pressure " , &pres ) ;

14
15
16
17
18
19

cout << " Potential \ tCurrent \ tTemperature \ tPressure \n" ;

for ( size_t irow =0; irow<in_chain . GetEntries ( ) ; ++irow ) {
in_chain . GetEntry ( irow ) ; // l o a d s a l l v a r i a b l e s t h a t h a v e

20
21
22

branches

cout

23

pot

<<

"\t"

<<

cur

<<

"\t"

<<

temp

<<

"\t"

<<

pres

<<

connected

to

endl ;

24
25

<<

been

}
le:

6.2.5.

For the advanced user:

read_ntuple_with_chain.cxx

Processing trees with a selector script

TChain::Process().
TSelector, and  optionally
 the number of entries and the rst entry to be processed. A template for the class TSelector is provided by
the method TTree::MakeSelector, as is shown in the little macro makeSelector.C below.
It opens the n-tuple conductivity_experiment.root from the example above and creates from it the header
le MySelector.h and a template to insert your own analysis code, MySelector.C.

Another very general and powerful way of processing a

TChain

is provided via the method

This method takes as arguments an instance of a  user-implemented class of type

33

6. File I/O

//

// / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / /

//

//

6
7

//

8
9

//

create

open

TFile

root
*

create

TTree
this

template

f
t

file

class

for

Selector

containing

the

to

run

on

tree

Tree

TFile : : Open ( " conductivity_experiment . root " ) ;

TTree

object

= ( TTree

generates

from

the

it

f>Get ( " cond_data " ) ;

*)

files

MySelector . h

10

t>MakeSelector ( " MySelector " ) ;

11

and

MySelector .C

le: makeMySelector.C
SlaveBegin() called before processing of the TChain starts,
Process() called for every entry of the chain, and SlaveTerminate() and Terminate() called after the last entry
has been processed. Typically, initialization like booking of histograms is performed in SlaveBegin(), the analysis,
i. e. the selection of entries, calculations and lling of histograms, is done in Process(), and nal operations like
plotting and storing of results happen in SlaveTerminate() or Terminate().
The entry points SlaveBegin() and SlaveTerminate() are called on so-called slave nodes only if parallel
processing via PROOF or PROOF lite is enabled, as will be explained below.
A simple example of a selector class is shown in the macro MySelector.C. The example is executed with the
The template contains the entry points

Begin()

and

following sequence of commands:

>

>

>

TChain * ch=new TChain ( " cond_data " , " My Chain for Example N - Tuple " ) ;
ch>Add ( " conductivity_experiment *. root " ) ;
ch>Process ( " MySelector .C+" ) ;

As usual, the  + appended to the name of the macro to be executed initiates the compilation of the

MySelector.C

with the system compiler in order to improve performance.

The code in

MySelector.C, shown in the listing below, books some histograms in SlaveBegin() and adds
fOutput, which is of the class TList1 The nal processing in Terminate() allows to access
store, display or save them as pictures. This is shown in the example via the TList fOutput. See

them to the instance

histograms and

the commented listing below for more details; most of the text is actually comments generated automatically by

TTree::MakeSelector.
1

# define

//

The

//

by

//

from

//

framework

//

The

//

//

10

//

11

//

12

//

13

//

14

//

15

//

16

//

17

//

18

//

19

/ / To

20

//

21

//

Root > T>P r o c e s s ( " M y S e l e c t o r . C" )

22

//

Root > T>P r o c e s s ( " M y S e l e c t o r . C" , " some

23

//

Root > T>P r o c e s s ( " M y S e l e c t o r . C+")

24

//

MySelector_cxx
class

definition

t h e ROOT u t i l i t y

in

MySelector . h

has

been

TTree : : M a k e S e l e c t o r ( ) .

t h e ROOT c l a s s

TSelector .

For

more

generated

This

class

information

s e e $ROOTSYS/README/README. SELECTOR o r

automatically

is

on

derived

the

TSelector

t h e ROOT U s e r

Manual .

following

methods

Begin ( ) :

are

a
SlaveBegin () :
Process () :

called

Terminate ( ) :

this

file ,

read

for
at

only

called

to

on

the

create
when

at

event ,

fill

the

your

end

on

the

convenient
the

file :

loop

Begin ( ) ,

each

and

called

try

this
a

place

after

called

in

time

tree

your

starts ,

histograms .

on PROOF c a l l e d

only

on

the

servers .

called
to

SlaveTerminate :

every

convenient

slave

use

defined

called

of

the
end

the

slave
of

place

following

in

the

to

this

function

you

decide

what

histograms .
loop
loop

draw / f i t

session

on

the

tree ,

when

on PROOF

servers .

on

on

the

your

your

tree ,
histograms .

Tree T:

options ")

25
26
27

# include " MySelector .h"

# include <TH2 . h>
1 The

usage of

fOutput is not really needed

PROOF (see next section).

processing with

34

for this simple example, but it allows re-usage of the exact code in parallel

6.2. N-tuples in ROOT

28
29

# include
# include

<T S t y l e . h>
<TCanvas . h>

30
31
32

//

user

defined

variables

may come

here :

UInt_t fNumberOfEvents ; TDatime tBegin , tNow ;

33
34

TH1F

h_pot , * h_cur , * h_temp , * h_pres , * h_resistance ;

35
36

void MySelector : : Begin ( TTree

37

38

//

39

/ / When r u n n i n g

The

Begin ( )

40

//

The

tree

function

/* t r e e */ )

is

called

at

w i t h PROOF B e g i n ( )

argument

is

deprecated

the

is

start

only

of

the

called

( on PROOF 0

on

is

query .
the

client .

passed ) .

41

TString option

42

GetOption ( ) ;

43

//

44

46
47

some

time

measurement

tBegin . Set ( ) ; printf ( " *==* ---------- Begin of Job ---------- " ) ;
tBegin . Print ( ) ;

45

48
49

void MySelector : : SlaveBegin ( TTree

50

51

//

52

/ / When r u n n i n g

The

SlaveBegin ()

53

//

The

tree

function

is

/* t r e e */ )
called

after

w i t h PROOF S l a v e B e g i n ( )

argument

is

deprecated

is

the

Begin ( )

called

( on PROOF 0

on

is

function .

each

slave

server .

passed ) .

54

TString option

55

GetOption ( ) ;

56

/ / book

57

some

histograms

h_pot=new TH1F ( " pot " , " potential " , 1 0 0 , 0 . 5 , 1 0 . 5 ) ;

h_cur=new TH1F ( " cur " , " current " , 1 0 0 , 0 . 1 , 1 . 5 ) ;
h_temp=new TH1F ( " temp " , " temperature " , 1 0 0 , 2 0 0 . , 4 0 0 . ) ;
h_pres=new TH1F ( " pres " , " pressure " , 1 0 0 , 0 . , 2 . ) ;
h_resistance=new TH1F ( " resistance " , " resistance " , 1 0 0 , 5 . , 1 5 . )

58
59
60
61
62

63

//

64
65
66

add

all

booked

histograms

to

output

list

( only

fOutput >AddAll ( gDirectory >GetList ( ) ) ;

really

needed

f o r PROOF)

67
68

Bool_t MySelector : : Process ( Long64_t entry )

69

70

//

The

71

//

keyed

Process ()

72

//

specifies

which

73

//

It

can

passed

74

//

to

read

75

//

keyed

76

//

via

77

//

78

//

This

79

//

simple

or

80

//

of

event

81

//

82

//

83

//

84

//

85

//

86

//

object
be

function
in

either

objects

the

entry
to

all

is

case
in

called
the

either
or

fObject

the

each

to

be

currently

entry

in

loaded

tree

M y S e l e c t o r : : GetEntry ( )
required
the

parts

object

is

the

processed .

of

the

already

tree

The
is

or

( or

entry

to

be

possibly

argument

processed .

TBranch : : G e t E n t r y ( )

data .
loaded

When

processing

and

is

available

pointer .

should

elaborate
and

contain
selection

typically

The

processing

can

be

Use

fStatus

set

the

The

return

to

for

o f PROOF)

w i t h PROOF,

function

the

the

value

is

the

fill

stopped
return

currently

" body "

of

criteria ,

run

the

analysis .

algorithms

on

It

can

the

contain

data

histograms .

by

calling

value
not

of

Abort ( ) .

TTree : : P r o c e s s ( ) .

used .

87
88
89

//

GetEntry ( entry ) ;

begin

processing

90

35

6. File I/O

//

91

count

number

of

entries

(= e v e n t s )

++fNumberOfEvents ;

92

...

93

//

94
95
96
97
98
99

a n a l s i y s c o d e comes h e r e f i l l h i s t o g r a m s
h_pot >Fill ( Potential ) ;
h_cur >Fill ( Current ) ;
h_temp >Fill ( Temperature ) ;
h_pres >Fill ( Pressure ) ;
h_resistance >Fill ( Potential / Current ) ;

100

return kTRUE ;

101
102

/ /kFALSE would

abort

processing

103
104

void MySelector : : SlaveTerminate ( )

105

106

//

The

107

//

have

SlaveTerminate ( )

108

//

on

been

each

processed .

slave

function

is

called

When r u n n i n g

after

all

entries

or

w i t h PROOF S l a v e T e r m i n a t e ( )

objects
is

called

server .

109

//

110

112
113
114
115
116
117

some

statistics

at

end

of

job

printf ( "\n *==* ---------- End of Slave Job ---------") ;

tNow . Set ( ) ; tNow . Print ( ) ;
printf ( " Number of Events : %i , elapsed time : %i sec , rate : %g evts / sec \n" ,
fNumberOfEvents ,
tNow . Convert ( )tBegin . Convert ( ) ,
float ( fNumberOfEvents ) / ( tNow . Convert ( )tBegin . Convert ( ) ) ) ;

111

118
119

void MySelector : : Terminate ( )

120

121

//

The

122

//

123

//

the

Terminate ( )

query .

It

results

function

always

is

the

runs

on

the

graphically

or

save

last

function

client ,
the

it

can

results

to

be

be

used

to

called
to

during

present

file .

124
125

//

126
127

finally ,

store

all

output

TFile hfile ( " MySelector_Result . root " , " RECREATE " , " MuonResults " ) ;
fOutput >Write ( ) ;

128

/ / Example

129

to

retrieve

output

from

output

list

h_resistance=dynamic_cast <TH1F * >( fOutput >FindObject ( " resistance " ) ) ;

TCanvas c_result ( " cresult " , " Resistance " , 1 0 0 , 1 0 0 , 3 0 0 , 3 0 0 ) ;
h_resistance >Draw ( ) ;
c_result . SaveAs ( " ResistanceDistribution . png " ) ;

130
131
132
133
134

tNow . Set ( ) ; printf ( " *==* ---------- End of Job ---------- " ) ;
tNow . Print ( ) ;

135
136
137

}
le:

6.2.6.

For power-users:

MySelector.C

Multi-core processing with PROOF lite

The processing of n-tuples via a selector function of type

TSelector through TChain::Process(), as described at

the end of the previous section, oers an additional advantage in particular for very large data sets: on distributed
systems or multi-core architectures, portions of data can be processed in parallel, thus signicantly reducing the
execution time.

On modern computers with multi-core CPUs or hyper-threading enabled, this allows a much

faster turnaround of analyses, since all the available CPU power is used.
On distributed systems, a PROOF server and worker nodes have to be set up, as is described in detail in the
ROOT documentation.
following little macro,

On a single computer with multiple cores,

RunMySelector.C,

the number of workers according to the number of CPU cores):

36

PROOF lite

can be used instead.

Try the

which contains two extra lines compared to the example above (adjust

6.2. N-tuples in ROOT

//

set

up

TChain

TChain * ch=new TChain ( " cond_data " , " My Chain for Example N - Tuple " ) ;
ch>Add ( " conductivity_experiment *. root " ) ;

//

//

7
8
9

eventually ,

start

Proof

Lite

TProof : : Open ( " workers =4 " ) ;

ch>SetProof ( ) ;

on

cores

//

10

ch>Process ( " MySelector .C+" ) ;

11

The rst command,

TProof::Open()

processing of the chain using PROOF. Now, when issuing the

code in

MySelector.C

is compiled and executed on each slave

are executed on the master only.

The list of n-tuple les is analysed, and portions of the data are assigned

to the available slave processes. Histograms booked in

and are lled accordingly.
merges them. In

SlaveBegin()

exist in the processes on the slave nodes,

Upon termination, the PROOF master collects the histograms from the slaves and

Terminate()

all merged histograms are available and can be inspected, analysed or stored. The

histograms are handled via the instances

this list after merging in

le: RunMySelector.C
ch->SetProof(); enables
command ch->Process("MySelector.C+);, the
node. The methods Begin() and Terminate()

starts a local PROOF server, and the command

Terminate.

fOutput

of class

TList

in each slave process, and can be retrieved from

To explore the power of this mechanism, generate some very large n-tuples using the script from Section 6.2.3
- you could try 10 000 000 events (this results in a large n-tuple of about 160 MByte in size).
generate a large number of les and use wildcards to add the to the

> root -l RunMySelector.C

TCHain.

You could also

Now execute

and watch what happens:

1
2
3
4
5

Processing RunMySelector . C . . .
+++ Starting PROOF Lite with 4 workers +++
Opening connections to workers : OK ( 4 workers )
Setting up worker servers : OK ( 4 workers )
PROOF set to parallel mode ( 4 workers )

6
7
8
9

10
11
12
13

14
15
16
17
18

Info in <TProofLite : : SetQueryRunning >: starting query : 1

Info in <TProofQueryResult : : SetRunning >: nwrks : 4
Info in <TUnixSystem : : ACLiC >: creating shared library ~/ DivingROOT / macros / MySelector_C . so
*==* Begin of Job Date / Time = Wed Feb 15 2 3 : 0 0 : 0 4 2 0 1 2
Looking up for exact location of files : OK ( 4 files )
Looking up for exact location of files : OK ( 4 files )
Info in <TPacketizerAdaptive : : TPacketizerAdaptive >: Setting max number of workers per node to 4
Validating files : OK ( 4 files )
Info in <TPacketizerAdaptive : : InitStats >: fraction of remote files 1 . 0 0 0 0 0 0
Info in <TCanvas : : Print >: file ResistanceDistribution . png has been created
*==* End of Job Date / Time = Wed Feb 15 2 3 : 0 0 : 0 8 2 0 1 2
Lite 0: all output objects have been merged
Log les of the whole processing chain are kept in the directory

~.proof

for each worker node. This is very

helpful for debugging or if something goes wrong. As the the method described here also works without using
PROOF, the development work on an analysis script can be done in the standard way on a small subset of the
data, and only for the full processing one would use parallelism via PROOF.

6.2.7. Optimisation Regarding N-tuples

ROOT automatically applies compression algorithms on n-tuples to reduce the memory consumption. A value
that is in most cases only zero will consume only small space on your disk (but it has to be deated on reading).
Nevertheless, you should think about the design of your n-tuples and your analyses as soon as the processing time
exceeds some minutes.

Try to keep your n-tuples simple and use appropriate variable types. If your measurement has only a limited
precision, it is needless to store it with double precision.

Experimental conditions that do not change with every single measurement should be stored in a separate
tree.

Although the compression can handle redundant values, the processing time increase with every

variable that has to be lled.

37

6. File I/O

The function

SetCacheSize(long)

species the size of the cache for reading a

TTree

object from a le.

The default value is 30MB. A manual increase may help in certain situations. Please note that the caching
mechanism can cover only one

TTree

object per

TFile

object.

You can select the branches to be covered by the caching algorithm with
unneeded branches with

SetBranchStatus.

AddBranchToCache and deactivate

This mechanism can result in a signicant speed-up for simple

operations on trees with many branches.

You can measure the performance easily with

TTreePerfStats. The ROOT documentation on this class

TTreePerfStats can show you that it is benecial to

also includes an introductory example. For example,

store meta data and payload data separately, i. e. write the meta data tree in a bulk to a le at the end of
your job instead of writing both trees interleaved.

38

CHAPTER 7
FUNCTIONS AND PARAMETER ESTIMATION

After going through the previous chapters, you already know how to use mathematical functions (class

TF1),

and you got some insight into the graph (TGraphErrors) and histogram classes (TH1F) for data visualisation. In
this chapter we will add more detail to the previous approximate explanations to face the fundamental topic
of parameter estimation by tting functions to data.

For graphs and histograms, ROOT oers an easy-to-

use interface to perform ts - either the t panel of the graphical interface, or the

TVirtualFitter

Fit

method.

The class

allows access to the detailed results, and can also be used for more general tasks with user-

dened minimisation functions.

Very often it is necessary to study the statistical properties of analysis procedures. This is most easily achieved
by applying the analysis to many sets of simulated data (or pseudo data), each representing one possible version
of the true experiment. If the simulation only deals with the nal distributions observed in data, and does not
perform a full simulation of the underlying physics and the experimental apparatus, the name Toy Monte Carlo

is frequently used . Since the true values of all parameters are known in the pseudo-data, the dierences between
the parameter estimates from the analysis procedure w. r. t.

the true values can be determined, and it is also

possible to check that the analysis procedure provides correct error estimates.

7.1. Fitting Functions to Pseudo Data

In the example below, a pseudo-data set is produced and a model tted to it.
ROOT oers various t methods, all inheriting from a virtual class

TVirtualFitter.

The default tter in

ROOT is MINUIT, a classical tting package originally implemented in the FORTRAN programming language.
Recently, a C++ version, MINUIT2, has been added, and the new package FUMILI. All of these methods
2
determine the best-t parameters, their errors and correlations by minimising a or a negative log-likelihood
function. A pointer to the active tting method is accessible via an instance of class

TVirtualFitter.

Methods of

this class allow to set initial values or allowed ranges for the t parameters, provide means for xing and releasing
of parameters and oer steering options for the numerical precision, and - most importantly - allow to retrieve
the status of the t upon completion and the t results. The documentation of the class

TVirtualFitter

gives a

list of all currently implemented methods.

The complexity level of the code below is intentionally a little higher than in the previous examples.

The

graphical output of the macro is shown in Figure 7.1:

/*

Define

and

play

with

TF1s

*/

2
3
4
5
6

void format_line ( TAttLine * line , int col , int sty ) {

line>SetLineWidth ( 5 ) ;
line>SetLineColor ( col ) ;
line>SetLineStyle ( sty ) ; }

7
8
9
10

double the_gausppar ( double * vars , double * pars ) {

return pars [ 0 ] * TMath : : Gaus ( vars [ 0 ] , pars [ 1 ] , pars [ 2 ] ) +
pars [ 3 ] + pars [ 4 ] * vars [ 0 ] + pars [ 5 ] * vars [ 0 ] * vars [ 0 ] ; }

11

Monte Carlo simulation means that random numbers play a role here which is as crucial as in games of pure chance in
the Casino of Monte Carlo.

39

7. Functions and Parameter Estimation

12
13
14
15
16
17
18

int macro8 ( ) {
gROOT >SetStyle ( " Plain " ) ;
gStyle >SetOptTitle ( 0 ) ;
gStyle >SetOptStat ( 0 ) ;
gStyle >SetOptFit ( 1 1 1 1 ) ;
gStyle >SetStatX ( . 8 9 ) ; gStyle >SetStatY ( . 8 9 )
gStyle >SetStatBorderSize ( 0 ) ;

19

TF1 parabola ( " parabola " , " [0]+[1]* x +[2]* x **2 " , 0 , 2 0 ) ;
format_line (& parabola , kBlue , 2 ) ;

20
21
22

TF1 gaussian ( " gaussian " , " [0]* TMath :: Gaus (x ,[1] ,[2]) " , 0 , 2 0 ) ;
format_line (& gaussian , kRed , 2 ) ;

23
24
25

TF1 gausppar ( " gausppar " , the_gausppar , 0 , 2 0 , 6 ) ;

double a = 1 5 ; double b = 1.2; double c = . 0 3 ;
double norm =4; double mean =7; double sigma =1;
gausppar . SetParameters ( norm , mean , sigma , a , b , c ) ;
gausppar . SetParNames ( " Norm " , " Mean " , " Sigma " , "a" , "b" , "c" ) ;
format_line (& gausppar , kBlue , 1 ) ;

26
27
28
29
30
31
32

TH1F histo ( " histo " ,

" Signal plus background ;X vals ;Y Vals " ,

33
34

50 ,0 ,20) ;

35

histo . SetMarkerStyle ( 8 ) ;

36
37

//

38

for

39
40

Fake

( int

the

data

i =1; i <=5000;++ i )
histo . Fill ( gausppar . GetRandom ( ) ) ;

41
42

/*

Reset

the

43

by

eye

peak

//

perform

fit

//

...

retrieve

parameters
at

with

before
an

area

the
of

45
46
47
48

and
or

set
50

*/

TVirtualFitter : : GetFitter ( ) ;

//

gausppar . SetParameter ( 0 , 5 0 ) ;
gausppar . SetParameter ( 1 , 6 ) ;
int npar=gausppar . GetNpar ( ) ;
for ( int ipar =2; ipar<npar ;++ ipar )
gausppar . SetParameter ( ipar , 1 ) ;

44

fit
more

less

49
50

...

histo . Fit (& gausppar ) ;

51
52
53

and

55

TVirtualFitter * fit
fit>PrintResults ( 2

56

//

54

get

covariance

fit
=

,0.) ;

//

print

fit

Matrix

an

print

it

TMatrixD * covMatrix
covMatrix >Print ( ) ;

57
58

results

get

fit

method

results

new TMatrixD ( npar , npar , fit>GetCovarianceMatrix ( ) ) ;

59

//

60

for

61
62
63

Set

the

( int

values

of

the

gaussian

and

parabola

ipar =0; ipar <3; ipar ++){

gaussian . SetParameter ( ipar , gausppar . GetParameter ( ipar ) ) ;
parabola . SetParameter ( ipar , gausppar . GetParameter ( ipar +3) ) ; }

64

histo . GetYaxis ( )>SetRangeUser ( 0 , 2 5 0 ) ;

histo . DrawClone ( " PE " ) ;
parabola . DrawClone ( " Same " ) ; gaussian . DrawClone ( " Same " ) ;
TLatex latex ( 2 , 2 2 0 , "# splitline { Signal Peak over }{ background }" ) ;
latex . DrawClone ( " Same " ) ;
return 0 ;

65
66
67
68
69
70
71

le:macro8.cxx

40

7.2. Toy Monte Carlo Experiments

Line 3-6:

A simple function to ease the make-up of lines.

TAttLine.

Remember that the class

TF1

inherits from

Line 8-10: Denition of a customised function, namely a Gaussian (the signal) plus a parabolic function,
the background.

Line 13-18: Some maquillage for the Canvas. In particular we want that the parameters of the t appear
very clearly and nicely on the plot.

Line 26-31: dene and initialise an instance of

Line 33-40: dene and ll a histogram.

Line 42-48:

TF1.

for convenience, the same function as for the generation of the pseudo-data is used in the

t; hence, we need to reset the function parameters. This part of the code is very important for each t
procedure, as it sets the initial values of the t.

Line 51: A very simple command, well known by now: t the function to the histogram.

Line 5358: retrieve the output from the t Here, we simply print the t result and access and print the
covariance matrix of the parameters.

Line 60end: plot the pseudo-data, the tted function and the signal and background components at the

Y Vals

best-t values.

250

2 / ndf
Prob
Norm
Mean
Sigma
a
b
c

Signal Peak over

background
200

150

42.03 / 44
0.5564
57.83 8.01
7.01 0.14
0.9238 0.1652
200.6 5.4
-16.73 1.03
0.4438 0.0461

100

50

0
0

Figure 7.1.:

10

12

14

16

18
20
X vals

Function t to pseudo-data

7.2. Toy Monte Carlo Experiments

Let us look at a simple example of a toy experiment comparing two methods to t a function to a histogram, the
2 method and a method called binned log-likelihood t, both available in ROOT.
As a very simple yet powerful quantity to check the quality of the t results, we construct for each pseudo-data
set the so-called pull, the dierence of the estimated and the true value of a parameter, normalised to the
estimated error on the parameter,

(pestim ptrue )/p .

If everything is OK, the distribution of the pull values is a

standard normal distribution, i. e. a Gaussian distribution centred around zero with a standard deviation of one.
The macro performs a rather big number of toy experiments, where a histogram is repeatedly lled with
Gaussian distributed numbers, representing the pseudo-data in this example.

Each time, a t is performed

according to the selected method, and the pull is calculated and lled into a histogram. Here is the code:

/*

3
4

Toy Monte
check

pull

Carlo

example

distribution

to

comp are

chi2

and

binned

l o g l i k e l i h o o d

methods

*/

void pull ( int n_toys

= 10000 ,

41

7. Functions and Parameter Estimation

int n_tot_entries
int nbins = 4 0 ,
bool do_chi2=true

5
6
7

= 100 ,
){

gROOT >SetStyle ( " Plain " ) ;

9
10

TString method_prefix ( "Log - Likelihood " ) ;

if ( do_chi2 )
method_prefix="# chi ^{2} " ;

11
12
13
14
15

//

Create

histo

//

Histogram

for

//

Make

canvases

TH1F * h4 = new TH1F ( method_prefix+" h4 " , method_prefix+" Random Gauss " , nbins , 4 ,4) ;
h4>SetMarkerStyle ( 2 1 ) ;
h4>SetMarkerSize ( 0 . 8 ) ;
h4>SetMarkerColor ( kRed ) ;

16

17
18
19
20
21

sigma

and

pull

TH1F * sigma = new TH1F ( method_prefix+" sigma " , method_prefix+" sigma from gaus fit " , 5 0 , 0 . 5 , 1 . 5 ) ;
TH1F * pull = new TH1F ( method_prefix+" pull " , method_prefix+" pull from gaus fit " ,50 , 4. ,4.) ;

22

23

24
25

nice

TCanvas * c0

26

new TCanvas ( method_prefix+" Gauss " , method_prefix+" Gauss " -

,0 ,0 ,320 ,240) ;

c0>SetGrid ( ) ;

27
28

//

29

Make

nice

canvases

TCanvas * c1 = new TCanvas ( method_prefix+" Result " , method_prefix+" Sigma - Distribution " , 0 , 3 0 0 , 6 0 0 , 4 0 0 ) ;

30

31

c0>cd ( ) ;

32
33

float sig , mean ;

for ( int i =0; i<n_toys ; i++){

34
35

//

36

Reset

//

38

Fill

40

//

perform

43

//

some

if

45
46
47

fit

control

output

i %100) ) {
h4>Draw ( " EP " ) ;
c0>Update ( ) ;

on

the

way

(!(

48

//

49

Get

51
52
53
54

55

//

56
57
58
59
60

62

42

sigma

from

fit

TF1 * fitfunc = h4>GetFunction ( " gaus " ) ;

sig = fitfunc >GetParameter ( 2 ) ;
mean= fitfunc >GetParameter ( 1 ) ;
sigma >Fill ( sig ) ;
pull>Fill ( mean / sig * sqrt ( n_tot_entries ) ) ;

50

61

histo

if ( do_chi2 ) h4>Fit ( " gaus " , "q" ) ; // C h i 2 f i t

else h4>Fit ( " gaus " , " lq " ) ; // L i k e l i h o o d f i t

42

44

contents

for ( int j = 0 ; j<n_tot_entries ; j++

h4>Fill ( gRandom >Gaus ( ) ) ;

39

41

histo

h4>Reset ( ) ;

37

//

end

print

of

t o y MC l o o p

result

c1>cd ( ) ;
pull>Fit ( " gaus " ) ;
pull>Draw ( " EP " ) ;
c1>Update ( ) ;

7.3. Fitting in General

63
64
65
66
67
68
69
70

void macro9 ( ) {
int n_toys = 1 0 0 0 0 ;
int n_tot_entries = 1 0 0 ;
int n_bins = 4 0 ;
cout << " Performing Pull Experiment with chi2 \n" ;
pull ( n_toys , n_tot_entries , n_bins , true ) ;
cout << " Performing Pull Experiment with Log Likelihood \n" ;
pull ( n_toys , n_tot_entries , n_bins , false ) ;
}

71

le:

macro9.cxx

Your present knowledge of ROOT should be enough to understand all the technicalities behind the macro.

pull in line 54 is dierent from the denition above: instead of the parameter error
mean, the tted standard deviation of the distribution divided by the square root of the number of entries,
sig/sqrt(n_tot_entries), is used.

Note that the variable

on

What method exhibits the better performance with the default parameters?

What happens if you increase the number of entries per histogram by a factor of ten? Why?

7.3. Fitting in General

In the examples above, we used the simplied tting interface of ROOT, and the default minimisation functions.
In general, however, tting tasks often require special, user-dened minimisation functions.

This is the case

when data cannot be represented as one- or two-dimensional histograms or graphs, when errors are correlated
and covariance matrices must be taken into account, or when external constrains on some of the t parameters
exist. The default minimiser in ROOT is MINUIT, a package that has been in use since decades. It oers several

TMinuit. A more modern,

TVirtualFitter), but still lacks some

minimisation methods and a large number of features accessible through the class
generalised interface allowing to use other minimises also exists (see class

TMinuit.

of the original features oered by

The macro below provides a rather general example. Data is read from a le, stored in an n-tuple for repeated
access, and an extended negative log-likelihood function is calculated and minimised with respect to the t
parameters.

/*

Example

of

based

on

t h e ROOT c l a s s e s

negative

log

likelihood

fit

( unbinned )

TVirtualFitter

and

TMinuit
*/

4
5
6
7
8

//

global

variables

TF1 * PDF ;
TNtuple * inpdata ;
int NFitPar =2;

for

this
//

macro
probability

/ / n t u p l e
//

to

specify

density

hold

number

function

input
of

for

the

fit

data

fit

parameters

9
10

/ /

11

/ / The

function

to

be

minimized ,

16

void fFCN ( Int_t &npar ,

double * gin ,
double &n2lL ,
double * par ,
int iflag )

17

12
13
14
15

//

18

20
21
22
23
24
25
26

//

set

negative

parameters

called

number

of

//

array

//

the

//

array

//

inernal

log

of

derivatives

function
of

by MINUIT ,

parameters ,

must

have

this

form .

optional
w. r . t .

parameters ,

optional

value

parameters
flag

likelihood

o f PDF

PDF>SetParameters ( par [ 0 ] , par [ 1 ] ) ;

// c a l c u l a t e l o g L , i . e . l o o p o v e r n t u p l e
float * ntrow ;
for ( int i =0; i < inpdata >GetEntries ( ) ; ++i ) {
inpdata >GetEntry ( i ) ; ntrow=inpdata >GetArgs ( ) ;
n2lL = log ( PDF>Eval ( ntrow [ 0 ] ) ) ;
}

27

n2lL

28
29

calculate

n2lL = 0 . ;

19

//

*=

2.;

// m u l t i p l y

by

two

( as

common

elsewhere

i n ROOT)

43

7. Functions and Parameter Estimation

30
31
32

main
void negLogLfit ( ) {

//

program ,

fit

control

33
34
35
36

//

define

probability

density

function ,

normalized

to

one

PDF=new TF1 ( " eplusconstPDF " ,

" (1. -[1]) *( exp (-x /[0]) -exp ( -5./[0]) ) /[0]+[1]/(5.) " ,
//

37

exponential

in

range

[0 ,5.]

0. ,

5.) ;

o f f s e t

plus

38
39
40
41
42
43
44
45

//

input

data

//

read

//

create

come

from

file

and

are

stored

in

an

NTuple

inpdata=new TNtuple ( " InputData " , " InputData " , "x" ) ;

data

from

file

and

store

in

ntuple

ifstream inp ; double x ;

inp . open ( " expob . dat " ) ;
while ( ! ( inp >> x ) ==0){ inpdata >Fill ( x ) ; }
inp . close ( ) ;

46
47
48
49
50

fitter

instance

and

initialize

( using

Minuit )

TVirtualFitter : : SetDefaultFitter ( " Minuit " ) ;

TFitter * fit =( TFitter * ) TVirtualFitter : : Fitter ( NULL , NFitPar ) ;
/* t h e above t y p e c a s t i s a b i t u g l y needed t o g a i n a c c e s s
TMinuit

51

object ,

which

is

possible

only

// a s s i g n

function

to

via

class

TFitter

to

*/

52
53

fit>SetFCN ( fFCN ) ;

be

minimized

54
55
56
57

//

set

initial

values

of

parameters

fit>SetParameter ( 0 ,
// p a r a m e t e r
" tau " , / / p a r a m e t e r name

58

1. ,

//

initial

value

59

0.1 ,

//

initial

uncertainty ,

60

0,

//

upper

0) ;

//

lower

61
62
63

//

limit
limit
name

fit>SetParameter ( 1 , " off " ,

index

fixes

0:

not

set

0:

not

set

val

err

0.5 ,

low

0.1 ,

parameter

up

0,

0) ;

64
65
66
67

//

run

the

fit

double arglist [ 2 ] = { 5 0 0 0 , 0 . 0 1 } ;
/ / {max .
fit>ExecuteCommand ( " MINIMIZE " , arglist ,

number

of

2) ;

function
//

calls ,

tolerance }

p e r f o r m s SIMPLEX + MIGRAD-

algorithms

fit>ExecuteCommand ( " MINOS " , arglist ,

68

0) ;

/ / MINOS

error

evaluation

69
70

//

71
72
73
74

75
76

//

77
78

r e t r i e v e o u t p u t
int nvpar , nparx ; double amin , edm , errdef ;
if ( fit>GetStats ( amin , edm , errdef , nvpar , nparx ) ==3){
cout<<endl<<" *==* Fit converged :"
<< " nlL ="<<amin <<" edm ="<<edm<<" nvpar ="<<nvpar <<" nparx ="<<nparx <<endl<<endl ;
fit>PrintResults ( 4 , amin ) ; }
get

covariance

Matrix

TMatrixD * covMatrix
covMatrix >Print ( ) ;

an

print

it

new TMatrixD ( nparx , nparx , fit>GetCovarianceMatrix ( ) ) ;

79
80

// p l o t

data ,

fit

result ,

and

parameter

contours

TCanvas * c = new TCanvas ( "c" , " contours " , 1 0 , 1 0 , 4 0 0 , 6 0 0 ) ;

c>Divide ( 1 , 2 ) ;
c>cd ( 1 ) ;
inpdata >Draw ( "x" ) ;
TH1F * htemp = ( TH1F * ) gPad>GetPrimitive ( " htemp " ) ;
TH1F * ht =( TH1F * ) htemp >Clone ( ) ;
ht>SetLineWidth ( 2 ) ;
ht>SetLineColor ( kBlue ) ;

81
82
83
84
85
86
87
88

/ / PDF must

89

be

scaled

to

take

account

of # of

Entries

ht>Eval ( PDF ) ; ht>Scale ( inpdata >GetEntries ( )

90

44

and

bin

width

ht>GetBinWidth ( 1 )

);

7.3. Fitting in General

ht>Draw ( "C SAME " ) ;

91
92
93

//

plot

contours

c>cd ( 2 ) ;

94
95

/ / Get

contour

for

parameter

versus

parameter

/ / Get

contour

for

parameter

versus

parameter

f o r ERRDEF=4

fit>SetErrorDef ( 4 ) ; // n o t e 4 and n o t 2 !
TMinuit * minuit= fit>GetMinuit ( ) ;
TGraph * gr2 = ( TGraph * ) minuit >Contour ( 4 0 , 0 , 1 ) ;
gr2>SetTitle ( " 1# sigma and 2# sigma contours ; tau ;off - set " ) ;
gr2>SetFillColor ( 4 2 ) ;
gr2>Draw ( " alf " ) ;

96
97
98
99
100
101
102

fit>SetErrorDef ( 1 ) ;
TGraph * gr1 = ( TGraph * ) minuit >Contour ( 4 0 , 0 , 1 ) ;
gr1>SetFillColor ( 3 8 ) ;
gr1>Draw ( " lf " ) ;

103
104
105
106

f o r ERRDEF=1

107

// c l e a n

110

up

delete inpdata ; delete PDF ;

109

negLogLfit.cxx

le:

You already know most of the code fragments used above. The new part is the user-dened minimisation function

fFCN,

made known to the minimiser via the method

SetFCN(void *f).

Lines 1129: denition of function to be minimised; the parameter list (number of parameters, eventually
analytically calculated derivatives w.r.t.

the parameters, the return value of the function, the array of

parameters, and a control ag) is xed, as it is expected by the minimisation package.

This function is

repeatedly called by the minimisation package with dierent values of the function parameters.

Lines 3563: initialisation of the t: denition of a probability density function as a

TF1, creation and lling

of an n-tuple containing the data read from a le, and the denition of the t parameters and their initial
values and ranges. The minimiser is instantiated in lines 48 and 49.

MINOS method.

Lines 6668 execute the t, rst a general minimisation, and then an error analysis using the

Lines 70106: retrieval of t results after completion of the t; this part needs access to the data and serves
for a comparison of the t result with the data - here, we show the tted function on top of a histogram of
the input data. Note that the PDF of a likelihood t needs to be scaled to take into account the bin width
of the histogram.

Line 6391: The function

printFit

illustrates how to access the best-t values of the parameters and their

errors and correlations from an object of

TMinuit.

Here, they are written to standard output; it is easy to

redirect this into a le or some other data structure, if required.

Code starting at line 93 illustrates how contour lines of two t parameters of one and two
The correlation of the two variables

tau

and

off-set

htemp

Entries
170
Mean
2.045
RMS
1.39
Underflow
0
Overflow
0

8
7
6

are produced.

is clearly visible (Figure 7.2).

Contours
off-set

108

0.95

0.9

0.85

5
0.8
4
0.75

0.7

1
0.65
0
0

0.4

0.5

0.6

0.7

0.8

0.9

1.1

1.2

Figure 7.2.: Histogrammed input data with overlayed scaled t function, and one- and 2- contour lines
from the likelihood t.

45

CHAPTER 8
ROOT IN

PYTHON

Python
Python is used in a wide variety of application areas and one of the most used scripting

ROOT also oers an interface named PyRoot, see http://root.cern.ch/drupal/content/pyroot, to the

programming language.

languages today. With its very high-level data types with dynamic typing, its intuitive object orientation and the
clear and ecient syntax

Python is very suited to control even complicated analysis work ows.

With the help

of PyROOT it becomes possible to combine the power of a scripting language with ROOT methods.
Introductory material to

Python is available from many sources in the Internet, see e. g. http://docs.python.org/.

Python packages, like numpy , providing high-level mathematical functions

There are additional very powerful

and handling of large multi-dimensional matrices, or

matplotlib ,

providing plotting tools for publication-quality

graphics. PyROOT additionally adds to this access to the vast capabilities of the ROOT universe.

Python, the environment variable PYTHONPATH must include the path to the library path,
Python support. Then, PyROOT provides direct interactions with
ROOT classes from Python by importing ROOT.py into Python scrips via the command import ROOT; it is
To use ROOT from

$ROOTSYS/lib,

of a ROOT version with

also possible to import only selected classes from ROOT, e. g.

from ROOT import TF1.

8.1. PyROOT
The access to ROOT classes and their methods in PyROOT is almost identical to C++ macros, except for the
special language features of

Python, most importantly dynamic type declaration at the time of assignment.

Coming back to our rst example, simply plotting a function in ROOT, the following C++ code:

1
2
3
4

TF1 * f1 = new TF1 ( " f2 " , " [0]* sin ([1]* x)/x"
f1>SetParameter ( 0 , 1 ) ;
f1>SetParameter ( 1 , 1 ) ;
f1>Draw ( ) ;
in

1
2
3
4
5

,0. ,10.) ;

Python becomes:

import ROOT
f1 = ROOT . TF1 ( " f2 " , " [0]* sin ([1]* x)/x"
f1 . SetParameter ( 0 , 1 )
f1 . SetParameter ( 1 , 1 )
f1 . Draw ( ) ;

,0. ,10.)

A slightly more advanced example hands over data dened in the macro to the ROOT class
Note that a

TGraphErrors.

Python array can be used to pass data between Python and ROOT. The rst line in the Python

script allows it to be executed directly from the operating system, without the need to start the script from

python

or the highly recommended powerful interactive shell

ipython.

The last line in the python script is there

to allow you to have a look at the graphical output in the ROOT canvas before it disappears upon termination
of the script.

47

8. ROOT in

Python

Here is the C++ version:

void TGraphFit ( )

//

/ / Draw a

//

graph

with

error

gStyle >SetOptFit ( 1 1 1 ) ;

//

make

nice

bars

and

fit

// s u p e r i m p o s e

function

fit

to

it

results

Canvas

TCanvas * c1 = new TCanvas ( " c1 " , " Daten " , 2 0 0 , 1 0 , 7 0 0 , 5 0 0 ) ;

c1>SetGrid ( ) ;

// d e f i n e

some

14

const Int_t n
Float_t x [ n ]
Float_t y [ n ]
Float_t ey [ n ]
Float_t ex [ n ]

15

//

and

hand

//

now

perform

10
11
12
13

16
17
18
19

data

= { 0.22 ,
= {0.7 ,

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

9. ,

0.5 ,

0.61 ,

9.6 ,

8.7 ,

0.7 ,
6.3 ,

0.85 ,
4.5 ,

over

to

TGraphErros

0.89 ,

1.1};

1.1};

object

TGraphErrors * gr = new TGraphErrors ( n , x , y , ex , ey ) ;

gr>SetTitle ( " TGraphErrors with Fit " ) ;
gr>Draw ( " AP " ) ;
a

fit

( with

errors

in

and

y!)

le:

In

0.35 ,

7.4 ,

{.8 ,.7 ,.6 ,.5 ,.4 ,.4 ,.5 ,.6 ,.7 ,.8};

0.25 ,

5.6 ,

{.05 ,.1 ,.07 ,.07 ,.04 ,.05 ,.06 ,.07 ,.08 ,.05};

22

0.1 ,

2.9 ,

21

...

gr>Fit ( " gaus " ) ;

c1>Update ( ) ;

20

points

= 10;

TGraphFit.C

Python it looks like this:

# ! / u s r / b i n / env python
#
# Draw a g r a p h w i t h error b a r s and f i t a f u n c t i o n t o i t
#
from ROOT import gStyle , TCanvas , TGraphErrors
from array import array
gStyle . SetOptFit ( 1 1 1 ) # superimpose fit results
c1=TCanvas ( " c1 " , " Data " , 2 0 0 , 1 0 , 7 0 0 , 5 0 0 ) #make nice Canvas
c1 . SetGrid ( )
# define some d a t a p o i n t s . . .
x = array ( 'f ' , ( 0 . 2 2 , 0 . 1 , 0 . 2 5 , 0 . 3 5 , 0 . 5 , 0 . 6 1 , 0 . 7 , 0 . 8 5 , 0 . 8 9 ,
y = array ( 'f ' , ( 0 . 7 , 2 . 9 , 5 . 6 , 7 . 4 , 9 . , 9 . 6 , 8 . 7 , 6 . 3 , 4 . 5 , 1 . 1 ) )
ey = array ( 'f ' , ( . 8 , . 7 , . 6 , . 5 , . 4 , . 4 , . 5 , . 6 , . 7 , . 8 ) )
ex = array ( 'f ' , ( . 0 5 , . 1 , . 0 7 , . 0 7 , . 0 4 , . 0 5 , . 0 6 , . 0 7 , . 0 8 , . 0 5 ) )
nPoints=len ( x )
# . . . and hand o v e r t o TGraphErros o b j e c t
gr=TGraphErrors ( nPoints , x , y , ex , ey )
gr . SetTitle ( " TGraphErrors with Fit " )
gr . Draw ( " AP " ) ;
gr . Fit ( " gaus " )
c1 . Update ( )
# r e q u e s t u s e r a c t i o n b e f o r e e n d i n g ( and d e l e t i n g g r a p h i c s window )
raw_input ( ' Press <ret > to end -> ' )

1.1)

le:

TGraphFit.py

Python versions in these two examples, it now should be clear how easy it is to
Python version.
As another example, let us revisit macro3 from Chapter 4. A straight-forward Python version relying on the

Comparing the C++ and

convert any ROOT Macro in C++ to a

ROOT class

1
2
3
4

TMath:

# ! / u s r / b i n / env python
#
( t h e f i r s t line a l l o w s e x e c u t i o n d i r e c t l y f r o m t h e l i n u x
#
# macro3 a s p y t h o n s c r i p t

48

shell )

8.1. PyROOT

5
6
7
8
9
10

# Author :
G . Quast
Oct . 2 0 1 3
# d e p e n d e n c i e s : PYTHON v2 . 7 , p y r o o t
# l a s t modified :
#
#
# * * * B u i l d s a p o l a r g r a p h i n a s q u a r e Canvas

11
12
13

from ROOT import TCanvas , TGraphPolar , TMath

from array import array

14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

rmin =0.
rmax = 6 . * TMath . Pi ( )
npoints =300
r=array ( 'd ' , npoints * [ 0 . ] )
theta=array ( 'd ' , npoints * [ 0 . ] )
e=array ( 'd ' , npoints * [ 0 . ] )
for ipt in range ( 0 , npoints ) :
r [ ipt ] = ipt * ( rmaxrmin ) / ( npoints 1 . )+rmin
theta [ ipt ]= TMath . Sin ( r [ ipt ] )
c=TCanvas ( " myCanvas " , " myCanvas " , 6 0 0 , 6 0 0 )
grP1=TGraphPolar ( npoints , r , theta , e , e )
grP1 . SetTitle ( "A Fan " )
grP1 . SetLineWidth ( 3 )
grP1 . SetLineColor ( 2 )
grP1 . Draw ( " AOL " )

30
31

raw_input ( ' Press <ret > to end -> ' )

8.1.1.

More

le:

macro3.py

Python- less ROOT

Python modules providing functionality similar to ROOT

Python code.

You may have noticed already that there are some

classes, which t more seamlessly into your
A more pythonic version of the above
provisoining of data to

1
2
3
4
5
6
7
8
9

TGraphPolar.

macro3

With the

would use a replacement of the ROOT class

math package, the part of the code becomes

TMath

for the

import math
from array import array
from ROOT import TCanvas , TGraphPolar
...

ipt=range ( 0 , npoints )
r=array ( 'd ' , map ( lambda x : x * ( rmaxrmin ) / ( npoints 1 . )+rmin , ipt ) )
theta=array ( 'd ' , map ( math . sin , r ) )
e=array ( 'd ' , npoints * [ 0 . ] )
...
Using the very powerful package

numpy and the built-in functions to handle numerical arrays makes the Python

code more compact and readable:

1
2
3
4
5
6
7

import numpy as np
from ROOT import TCanvas , TGraphPolar
...

r=np . linspace ( rmin , rmax , npoints )

theta=np . sin ( r )
e=np . zeros ( npoints )
...

le:

Customised Binning
This example combines comfortable handling of arrays in
togram.

macro3_numpy.py

Python to dene variable bin sizes of a ROOT his-

All we need to know is the interface of the relevant ROOT class and its methods (from the ROOT

documentation):

49

8. ROOT in

TH1F ( const char * name , const char * title , Int_t nbinsx , const Double_t * xbins )

Here is the
1
2
3
4
5
6
7
8

Python

Python code:

import ROOT
from array import array
arrBins = array ( 'd ' , ( 1 , 4 , 9 , 1 6 ) ) # array of bin edges
histo = ROOT . TH1F ( " hist " , " hist " , len ( arrBins ) 1 , arrBins )
# f i l l i t w i t h e q u a l l y s p a c e d numbers
for i in range ( 1 , 1 6 ) :
histo . Fill ( i )
histo . Draw ( )

A t example in

Python using TMinuit from ROOT

One may even wish to go one step further and do most of the implementation directly in

le:

histrogram.py

Python, while using

TMinuit is used as the minimizer in a 2 -t.

2
and the -function are dened in Python and

only some ROOT classes. In the example below, the ROOT class
Data are provided as
iteratively called by

Python arrays, the function to be tted

The results are extracted to Python objects,

Minuit.

powerful and versatile python package

1
2
3
4
5
6
7
8
9
10
11
12
13
14

matplotlib.

and plotting is done via the very

# ! / u s r / b i n / env python
#
# p y t h o n s c r i p t
# EXAMPLE s h o w i n g how t o s e t up a f i t w i t h MINUIT using p y r o o t
#
# Author :
G . Quast
May 2 0 1 3
# d e p e n d e n c i e s : PYTHON v2 . 7 , p y r o o t , numpy , m a t p l o t l i b , a r r a y
# l a s t m o d i f i e d : Oct . 6 , 2 0 1 3
#
#
from ROOT import TMinuit , Double , Long
import numpy as np
from array import array as arr
import matplotlib . pyplot as plt

15

17

# > define some d a t a

ax = arr ( 'f ' , ( -

18

ay

19

ey

20

nPoints

16

0.05 ,0.36 ,0.68 ,0.80 ,1.09 ,1.46 ,1.71 ,1.83 ,2.44 ,2.09 ,3.72 ,4.36 ,4.60)

arr ( 'f ' ,

0.35 ,0.26 ,0.52 ,0.44 ,0.48 ,0.55 ,0.66 ,0.48 ,0.75 ,0.70 ,0.75 ,0.80 ,0.90)

arr ( 'f ' ,

len ( ax )

21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

# > S e t p a r a m e t e r s and f u n c t i o n t o f i t
# a l i s t w i t h c o n v e n i e n t names ,
name = [ "a" , "m" , "b" ]
# the i n i t i a l values ,
vstart = arr ( 'd ' , ( 1 . 0 , 1 . 0 , 1 . 0 ) )
# and t h e i n i t i a l s t e p s i z e
step =
arr ( 'd ' , ( 0 . 0 0 1 , 0 . 0 0 1 , 0 . 0 0 1 ) )
npar =len ( name )
#
# this d e f i n e s t h e f u n c t i o n we want t o f i t :
def fitfunc ( x , npar , apar ) :
a = apar [ 0 ]
m = apar [ 1 ]
b = apar [ 2 ]
f = Double ( 0 )
f=a * x * x + m * x + b
return f
#

50

0.06 ,0.07 ,0.05 ,0.05 ,0.07 ,0.07 ,0.09 ,0.10 ,0.11 ,0.10 ,0.11 ,0.12 ,0.10)
=

8.1. PyROOT

40
41
42
43
44
45
46
47
48
49
50

# > this i s t h e d e f i n i t i o n o f t h e f u n c t i o n t o m i n i m i z e , h e r e
def calcChi2 ( npar , apar ) :
chisq = 0 . 0
for i in range ( 0 , nPoints ) :
x = ax [ i ]
curFuncV = fitfunc ( x , npar , apar )
curYV = ay [ i ]
curYE = ey [ i ]
chisq += ( ( curYV curFuncV ) * ( curYV curFuncV ) ) /
return chisq

c h i ^2 f u n c t i o n

( curYE * curYE )

51
52
53
54
55
56
57
58
59
60

61
62
63

# t h e f u n c t i o n f c n c a l l e d by MINUIT r e p e a t e d l y w i t h v a r y i n g p a r a m e t e r s
#
NOTE: t h e f u n c t i o n name i s s e t v i a TMinuit . SetFCN
def fcn ( npar , deriv , f , apar , iflag ) :
""" meaning of parametrs :
npar :
number of parameters
deriv : aray of derivatives df / dp_i (x) , optional
f:
value of function to be minimised ( typically chi2 or negLogL )
apar :
the array of parameters
iflag : internal flag : 1 at first call , 3 at the last , 4 during minimisation
"""
f [ 0 ] = calcChi2 ( npar , apar )
#

64
65
66
67
68
69
70
71

# > s e t up MINUIT
myMinuit = TMinuit ( npar ) # initialize TMinuit with maximum of npar parameters
myMinuit . SetFCN ( fcn )
# set function to minimize
arglist = arr ( 'd ' , 2 * [ 0 . 0 1 ] ) # set error definition
ierflg = Long ( 0 )
arglist [ 0 ] = 1 .
# 1 sigma is Delta chi ^2 = 1
myMinuit . mnexcm ( " SET ERR " , arglist , 1 , ierflg )

72
73
74
75
76
77
78

# > S e t s t a r t i n g v a l u e s and s t e p s i z e s for p a r a m e t e r s

for i in range ( 0 , npar ) :
# Define the parameters for the fit
myMinuit . mnparm ( i , name [ i ] , vstart [ i ] , step [ i ] , 0 , 0 , ierflg )
arglist [ 0 ] = 6 0 0 0 # Number of calls to FCN before giving up .
arglist [ 1 ] = 0 . 3 # Tolerance
myMinuit . mnexcm ( " MIGRAD " , arglist , 2 , ierflg ) # execute the minimisation

79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95

# > c h e c k TMinuit s t a t u s
amin , edm , errdef = Double ( 0 . ) , Double ( 0 . ) , Double ( 0 . )
nvpar , nparx , icstat = Long ( 0 ) , Long ( 0 ) , Long ( 0 )
myMinuit . mnstat ( amin , edm , errdef , nvpar , nparx , icstat )
# me an in g o f p a r a m e t e r s :
#
amin : v a l u e o f f c n a t minimum (= c h i ^ 2 )
#
edm :
e s t i m a t e d d i s t a n c e t o mimimum
#
e r r d e f : d e l t a _ f c n u s e d t o define 1 s i g m a e r r o r s
#
n v p a r : number o f v a r i a b l e p a r a m e t e r s
#
n p a r x : t o t a l number o f p a r a m e t e r s
#
i c s t a t : s t a t u s o f error m a t r i x :
#
3= a c c u r a t e
#
2= f o r c e d p o s . d e f
#
1= a p p r o x i m a t i v e
#
0= not c a l c u l a t e d
myMinuit . mnprin ( 3 , amin ) # print out by Minuit

96
97
98
99
100
101

# > g e t r e s u l t s f r o m MINUIT
finalPar = [ ]
finalParErr = [ ]
p , pe = Double ( 0 ) , Double ( 0 )
for i in range ( 0 , npar ) :

51

8. ROOT in

myMinuit . GetParameter ( i , p , pe )
finalPar . append ( float ( p ) )
finalParErr . append ( float ( pe ) )

102
103
104
105
106
107
108

Python
#

retrieve parameters and errors

# get c o v a r i a n c e matrix
buf = arr ( 'd ' , npar * npar * [ 0 . ] )
myMinuit . mnemat ( buf , npar ) # retrieve error matrix
emat=np . array ( buf ) . reshape ( npar , npar )

109
110
111
112
113
114
115
116
117
118
119

# > p r o v i d e f o r m a t t e d o u t p u t o f r e s u l t s
print "\n"
print " *==* MINUIT fit completed :"
print ' fcn@minimum = %.3 g ' %( amin ) , " error code =" , ierflg , " status =" , icstat
print " Results : \t
value
error
corr . mat ."
for i in range ( 0 , npar ) :
print '
%s: \t %10.3 e +/ - %.1 e
'%( name [ i ] , finalPar [ i ] , finalParErr [ i ] ) ,
for j in range ( 0 , i ) :
print '%+.3 g ' %( emat [ i ] [ j ] / np . sqrt ( emat [ i ] [ i ] ) / np . sqrt ( emat [ j ] [ j ] ) ) ,
print ' '

120
121
122
123
124
125
126
127
128
129
130

# > p l o t r e s u l t using m a t p l o t l i b
plt . figure ( )
plt . errorbar ( ax , ay , yerr=ey , fmt="o" , label='data ' ) # the data
x=np . arange ( ax [ 0 ] , ax [ nPoints 1 ] , abs ( ( ax [ nPoints 1] ax [ 0 ] ) / 1 0 0 . ) )
y=fitfunc ( x , npar , finalPar ) # function at bestfitpoint
plt . title ( " Fit Result " )
plt . grid ( )
plt . plot ( x , y , label='fit function ' )
plt . legend ( loc =0)
plt . show ( )
le: fitting-example.py

52

CONCLUDING REMARKS

This is the end of our guided tour through ROOT for beginners. There is still a lot coming to mind to be said, but

ROOT home page

ROOT reference guide with the documentation of all ROOT classes, or the ROOT users guide.

by now you are experienced enough to use the ROOT documentation, most importantly the
and the

A very useful way for you to continue exploring ROOT is to study the examples in the sub-directory

tutorials/

of any ROOT installation.

There are some powerful additions to ROOT, e. g. packages named RooFit and RooStats providing a frame
work for model building, tting and statistical analysis.

The ROOT class

TMVA

oers multi-variate analysis

tools including an articial neural network and many other advanced methods for classication problems. The
remarkable ability of ROOT to handle large data volumes was already mentioned in this guide, implemented
through the class

TTree.

But there is still much more for you to explore ...

End of this guide ... but hopefully not of your interaction with ROOT !

53

APPENDIX A
ROOFILAB

A.1. Root-based tool for tting: RooFiLab

Although simple in principle, the fomulation of a problem in

C++

and the complex environment of the ROOT

framework pose a relativly high hurdle to overcome for the beginner. A simplication and extension of avialable
standard methods for function tting to one-dimesional distributions is the package RooFiLab (Root Fits for
Laboratory courses).

Based on ROOT, this program developed at KIT (Karlsruhe Institute of Technology,

URL http://www-ekp.physik.uni-Karlsruhe.de/ quast/RooFiLab) oers an easy-to-use, structured graphical user

interface and an ASCII input format for typical use cases in student laboratory courses.
both the x- and y-coordinate are also supported.

Correlated erros on

In the most general case, covariance matrices of the x- and

y-coordinates can be specied. There is also a simplied possibility for special cases of fully correlated absolute
or relative errors on the measurements. An example t is shown in Figure A.1.

Title of Graphic
0.12

0.1

Y Axis

0.08

0.06

0.04

Graph 1
Function

0.02

0
0

10

ab

FiL

oo

Figure A.1.:

X Axis

Example of a straight-line t with independent and correlated (systematic) errors on both

the x- and y-directions.

High exibility in the denition of the model is achieved by direct usage of the ROOT interpreter, which has
been extended to use named parameters instead of parameter numbers. In addition, more complex models can
be implemented as C or C++ functions, wich are compiled and linked at run-time.
The elements of the gracal user interface (see Figure A.2) and control via the input le are described in the
manual (le

RooFiLab.pdf

in the subdirectory

RooFiLab/doc,

in German language). A brief overview is given

here.

55

A. RooFiLab

A.1.1. Installation
RooFiLab

is availalbe, fully installed along with ROOT in a virtual machine

based on the Ubuntu distribution.

The compressed disk image is most easily imported into the freely available virtualisation tool

VirtualBox

for the

most common Linux distributions, for Windows versions XP and later and for Macintosh operating systems.
The program code of

RooFiLab is distributed from the URL given above as a compressed archive RooFiLab.tar.gz.

After unpacking, the installation under Linux proceeds by executing

make;

the le

Makefile

contains all necces-

sary instructions. A ROOT installation must be present and initialized, i.e. the environment variable
contain the path to the ROOT executable and

A.1.2. Usage of
RooFiLab

LD_LIBRARY_PATH

PATH

must

must point to the ROOT libraries.

RooFiLab

oers two windows: one is used for control, the other is for graphics output. The control window, as

depicted in FigureA.2, is separated into four Shutters, oering the following actions

data input and denition of functions and parameters

xing of start values and Fit-by-Eye

execution of the t, eventually iteratively by xing some of the free parameters

options for graphical output

Figure A.2.:

The gracal user interface of RooFiLab.

During execution, ROOT functionality is also available. Of particular importance are procedures for interactive
manilulations of the output graphcis and their export. As usual, the context menu is opened by right-klicking of
the components of the graph or via the Toolbar at the top of the graphics window.
In addition to interactive usage of the controls of the graphical interface, ts can also be executed automatically
by specication of control options in the input le denig the data inputs. After an interactive t, options can
thus be archived in the input le and then be used for repeated, automated ts.

A.2. Examples with RooFiLab

The following subsections show simple examples illustrating the usage of
for own applications.

1 http://www-ekp.physik.uni-karlsruhe.de/~quast/VMroot

56

RooFiLab

and may serve as the basis

A.2. Examples with

RooFiLab

A.2.1. Straight-line t with correlated erros in x and y

This

RooFiLab

input le contains several control lines and documents the available options.

comment lines starting with

#!

followed by a keyword.

The control command

Control lines are

#! dofit = true

triggers an

automated t dened by the input data and the control options in the le.

# straight-line fit to data with errors in x and y, incl. simple correlations

# ===========================================================================
#! staterrors = xy
#! systerrors = 0.02 0.04 rel rel
#! fit = "m*x+b" "m,b" "roofilab.fit"
#! initialvalues = 0.015 0
### command to execute fit
#! dofit = true
### show systematic erros as second error bar
#! secondgraph = syst
#!
#!
#!
#!
#!

title = "Fit to data with correlated errors"

graphlegend = "Data" bottom right
functionlegend = "Model" bottom right
xaxis = "X-values"
yaxis = "Y-values or f(x)"

#!
#!
#!
#!
#!

markersettings = 1.5 4 24
functionsettings = 1 3 2
grid = y
logscale = 0
savegraphic = "roofilab.eps"

# =================Eingabe der Daten ===================================

# values in up to four columns separated by whitespaces
#
(except for linebreaks or linefeeds)
# x
y
ex
ey
4.05 0.035 0.12 0.006
4.36 0.056 0.13 0.007
4.68 0.052 0.09 0.005
4.80 0.044 0.09 0.005
5.09 0.048 0.14 0.007
5.46 0.055 0.14 0.007
5.71 0.066 0.17 0.009
5.83 0.048 0.21 0.011
6.44 0.075 0.22 0.011
8.09 0.070 0.28 0.014
8.72 0.097 0.32 0.016
9.36 0.080 0.37 0.018
9.60 0.120 0.39 0.020

57

A. RooFiLab

A.2.2. Averaging correlated measurements

Averaging correlated measurements formally corresponds to a t of a constant. The measurements in this example
are the individual measurements of the mass of the Z Boson at the electron-positron collider LEP at CERN. The
common error of 1.7 MeV results from uncertainties in the centre-of-mass energy of the accelerator.

#! systerrors = 0 0.0017 abs abs

species this common aboslute error on each measurement.

# Mesurements of Z-Mass by AELPH, DELPHI, L3 and OPAL

# --------------------------------------------------# graphics options
#! markersettings = 1.5 4 24
#! functionsettings = 1 3 3
#! grid = y
# logscale = 0
# savegraphic = "roofilab.eps"
# saverfl = "data.rfl"
# plot lables
#! title = "averaging measurements"
#! xaxis = "n"
#! yaxis = "Mass of Z boson"
#! graphlegend = "Z mass measurements" bottom right
#! functionlegend = "average Z mass" bottom right
# fit control
#! fit = "m" "m" "average.fit"
#! initialvalues = 91.2
#! dofit = true
#! staterrors = y # control-command
#! systerrors = 0 0.0017 abs abs
# the data, LEP electroweak working group, CERN 2000
1 91.1893 0.0031
2 91.1863 0.0028
3 91.1894 0.0030
4 91.1853 0.0029

58

The line

A.2. Examples with

RooFiLab

A.2.3. Fit of a polynomyal to data with Poisson errors

This example show the t of a fourth-order polynomial to data with uncorrelated, Poissonian errors, i. e. erros
2
given by the square root of the data points. Although the errors are non-Gaussion in this case, a -t often
results in acceptable results. With the option

#! fitmethod = likelihood

a likelihood method can be selected.

In this case, the statistical errors are ignored and may be ommitted. For technical reasons, the x-values must be
equi-distant in this case (due to usage of ROOT-class

TH1).

##########################################################
#
example: fit of an angular distribution
##########################################################
# plot commands
#! title = "angular distribution "
#! xaxis = "cos(theta)"
#! yaxis = "number of events"
#! graphlegend ="observed rate " top left
#! functionlegend ="fitted cos(theta) distribution " top left
#! markersettings = 1.5 2 5
#! functionsettings = 1 3 3
# fit control
#! fit = "a4*x^4+a3*x^3+a2*x^2+a1*x+a0" "a0,a1,a2,a3,a4" "v_vs_cost.fit"
#! dofit = true
# fitmethod = likelihood # uncomment to perform a Log Likelihood fit
# definition of data
#! staterrors = y
# cost N
sqrt(N)
-0.9
81.
9.0
-0.7
50.
7.1
-0.5
35.
5.9
-0.3
27.
5.2
-0.1
26.
5.1
0.1
60.
7.7
0.3
106.
10.3
0.5
189.
13.7
0.7
318.
17.8
0.9
520.
22.8

59

A. RooFiLab

A.2.4. Correlated measurements with full covariance matrix

As a more complex example the averaging procedure for measurements of the W Boson mass is shown here.
Measurements of the four LEP experiments in two nal states have dierent systematic errors, which are correlated
among groups of measurements. These are specied in the full 88 covariance matrix, which is composed of 44
block matrices.
and the matrix

The control line

wmass.cov

#! covmatrices = 0 wmass.cov

species that not covariance matrix in

are to be used in the t.

# Mesurements of W-Mass by AELPH, DELPHI, L3 and OPAL

# --------------------------------------------------# ### example of fit with covariance matrix#
# --- graphics options
#! markersettings = 1.5 4 24
#! functionsettings = 1 3 3
#! grid = y
#! title = "averaging measurements"
#! xaxis = "n"
#! yaxis = "Mass of W boson"
#! graphlegend = "W mass measurements" top right
#! functionlegend = "average W mass" top right
# --- fit control
#! fit = "m" "m" "Wmittelung.fit"
#! initialvalues = 80.5
#! dofit = true
# --- the data (LEP electroweak working group, CERN 2006)
#! staterrors = 0
#! systerrors = 0 0 abs abs
#! covmatrices = 0 wmass.cov
1 80.429 0.059 # qqlv ALEPH
2 80.340 0.076 # qqlv DELPHI
3 80.213 0.071 # qqlv L3
4 80.449 0.062 # qqlv OPAL
5 80.475 0.082 # qqqq ALEPH
6 80.310 0.102 # qqqq DELPHI
7 80.323 0.091 # qqqq L3
8 80.353 0.081 # qqqq OPAL
//file wmass.cov
0.003481 0.000316
0.000316 0.005776
0.000316 0.000316
0.000316 0.000316
0.000383 0.000383
0.000383 0.000383
0.000383 0.000383
0.000383 0.000383

60

0.000316
0.000316
0.005041
0.000316
0.000383
0.000383
0.000383
0.000383

0.000316
0.000316
0.000316
0.003844
0.000383
0.000383
0.000383
0.000383

0.000383
0.000383
0.000383
0.000383
0.006724
0.001741
0.001741
0.001741

0.000383
0.000383
0.000383
0.000383
0.001741
0.010404
0.001741
0.001741

0.000383
0.000383
0.000383
0.000383
0.001741
0.001741
0.008281
0.001741

0.000383
0.000383
0.000383
0.000383
0.001741
0.001741
0.001741
0.006561

APPENDIX B
MARKERS, COLOURS, SYMBOLS

B.1. Colour Wheel and Graph Markers

+3

te
hi
kW
kB
la
ck

+3

+4

-9
+1 -8
+2 -7
+3 -6
+4 -5
+5 -4
+6 -3
+7 -2
+8 -1
+9 0
0
+1 eal
kT

+2

+1 -9
+2 -8
+3 -7
+4 -6
+5 -5
+6 -4
+7 -3
+8 -2
+9 -1
+10 0
kAzure

-9

-7

-4

-8

-6

-3

0
kC
ya

-10

+1
n

-2
+2

+3

-5
-1
+4

ROOT Color Wheel

Figure B.1.:

+1

-3

ed

-6

-4

-7

-9

+1
+2 -9
+3 -8
+4 -7
+5 -6
+6 -5
+7 -4
+8 -3
+9 -2
+1 -1
0
0
kV
iol
et

kGray

+1

kR

+2

-2

-10

-8

-5

-9

-1

-6

-7

-4
0

kMagenta

kS
pr
ing
0
+1
0
1
-1
-2 -3
-4
+
9
-2
-3 +8
-5
-6
-7
-4 +7
-5 +6
-8
-9
-6 +5
-10 -7 +4
-8 +3
-9 +2
+1

0
+1
+2

kGreen

-1

+3

ink
+4 +3 +2 +1 0
kP 10
+
0 +9
-1
-2 -3
-4
-1 +8
-2 +7 -5
-6
-7
-3 +6
-9
-4 +5 -8
-5 +4
-6 +3 -10
-7 +2
-8 +1
-9

+4

+3

kOrange
ow
ell
0 +10 +4
kY
0
-1 +9
+1
-2
+8 -1
4
2
+
-3 +7
-3
-4
+6 -5
-7
-2
-5 +5
-6
-9 -6 +4 -8
-5
-7 +3
-8
-10
-10 -8 +2
-9 +1

-2

-3
+1

+2

+3

+4

lue
kB

The wheel shows all available colours in ROOT and the codes to specify them and The
markers provided by ROOT.

Table B.1.: Alternative symbols to select the ROOT markers for graphs.

Integer

Description

Literal

Integer

Description

Literal

dot

kDot

21

full square

kFullSquare

kPlus

22

full triangle up

kFullTriangleUp

kStar

23

full triangle down

kFullTriangleDown

kCircle

24

open circle

kOpenCircle

kMultiply

25

open square

kOpenSquare

small dot

kFullDotSmall

26

open triangle up

kOpenTriangleUp

medium dot

kFullDotMedium

27

open diamond

kOpenDiamond

large scalable dot

kFullDotLarge

28

open cross

kOpenCross

20

full circle

kFullCircle

29

open star

kOpenStar

61

B. Markers, Colours, Symbols

B.2. Lines and Arrows

"----|>"
"<|----"
"<|----|>"
"--|>--"
"--<|--"
"---->"
"<----"
"<---->"
"-->--"
"--<--"

Figure B.2.: The arrows styles available in ROOT.

B.3. Latex Symbols

Lower case

#club

#diamond

#voidn

#aleph

#leq

#geq

#heart
#Jgothic

#LT

Upper case

Variations

alpha :

Alpha :

beta :

Beta :

#spade

gamma :

Gamma :

#Rgothic

delta :

Delta :

epsilon :

Epsilon :

#GT

#approx

#neq

#equiv

#propto

zeta :

Zeta :

#in

#notin

#subset

#notsubset

eta :

Eta :

theta :

Theta :

iota :

Iota :

kappa :

Kappa :

lambda :

Lambda :

Mu :

#supset

#subseteq

#supseteq

#oslash

#cap

#cup

#wedge

#vee

#ocopyright

#copyright

#oright

#void1

#trademark

#void3

#AA

#aa

mu :

#times

#divide

#/

nu :

#pm
#3dots

Nu :

#upoint

xi :

Xi :

#nabla

#partial

omicron :

Omicron :

#downleftarrow #corner

pi :

Pi :

#bullet

#circ

#infty

#voidb

varepsilon :

vartheta :

varsigma :

#doublequote #angle

#lbar

#cbar

#topbar

#ltbar

rho :

Rho :

#arcbottom

#arcbar

sigma :

Sigma :

#uparrow

#rightarrow

tau :

Tau :

#oplus

#surd

upsilon :

Upsilon :

varUpsilon :

#Uparrow

#Rightarrow

phi :

Phi :

varphi :

chi :

Chi :

psi :

Psi :

omega :

Omega :

varomega :

#downarrow

#arctop

#leftarrow

#leftrightarrow #otimes
#Downarrow

#Leftarrow

#Leftrightarrow #prod

#void8

h #hbar

#Box
#parallel

#sum
#perp

#bottombar

#int
#odot

Figure B.3.: The main Latex symbols that can be interpreted by the TLatex class.

62

APPENDIX C
MOST RELEVANT CLASSES AND THEIR METHODS

This list of classes and methods shows the most relevant ones, which have been considered in this guide. It is an
excerpt from the ROOT class reference guide.

TGraphErrors:

the graph class with error bars

create Graph frm le
TGraphErrors(const char* filename, const char* format = "%lg %lg %lg %lg", Option_t* option = "")
create graph fom C-arrays
TGraphErrors(Int_t n, const Float_t* x, const Float_t* y, const Float_t* ex = 0, const Float_t* ey = 0)
create graph from histogram
TGraphErrors(const TH1* h)
t a function
.Fit(TF1* f1, Option_t* option = "", Option_t* goption = "", Axis_t xmin = 0, Axis_t xmax = 0)

.Fit(const char* formula, Option_t* option = "", Option_t* goption = "", Axis_t xmin = 0, Axis_t xmax = 0)
.Draw("AP") and .DrawClone("AP")
methods of classes TGraph, TGraphPainter

draw
draw options

TH1F: the histogram class with oat bin contents

create (book) histogram
store also squared weights
ll a value
ll with weight
set bin content
get bin content
ll with random numbers
clear
copy to C-array
set maximum on y-axis
set minimum on y-axix
get mean
get RMS
draw
useful draw options

TH1F(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup)
.Sumw2()
.Fill(Double_t x)
.Fill(Double_t x, Double_t w)
.SetBinContent(Int_t bin, Double_t content)
Double_t .GetBinContent(Int_t bin) const
.FillRandom(const char* fname, Int_t ntimes)
.Reset()
Float_t* .GetArray()
.SetMaximum(Double_t ymax)
.SetMinimum(Double_t ymin)
Double_t GetMean(1)
Double_t GetRMS(1)
.Draw(Option_t* option = "")

"SAME" "E" "P"

see documentation of class THistPainter

TH2F: 2-dimensional histogram class with oat bin contents

book
ll
ll with weight
get mean along axis i
get RMS along axis i
get covariance
get correlation
draw
useful draw options

TProle:

TH2F(const char* name, const char* title, Int_t nbinsx, Double_t xlow, Double_t xup, Int_t nbinsy, Double_t ylow, Double_t yup)
Fill(Double_t x, Double_t y)
Fill(Double_t x, Double_t y, Double_t w)
Double_t GetMean(i)
Double_t GetRMS(i)
Double_t GetCovariance()
Double_t GetCorrelationFactor()
.Draw(Option_t* option = "") and .DrawClone
"" "SAME" "BOX" "COL" "LEGO" "SURF"
see documentation of class THistPainter

"prole representation" for 2-dim histograms

book prole histogram
TProfile(const char* name,const char* title,Int_t nbinsx,Double_t xlow,Double_t xup,Double_t ylow,Double_t yup,Option_t* option = "")
ll a value
.Fill(Double_t x)
ll with weight
.Fill(Double_t x, Double_t w)
draw
.Draw() and .DrawClone()

TF1:

the mathematical function

dene function in TFormula syntax
predened functions
dene function via pointer
evaluate at x
calculate derivative
calculate integral a to b
get random number
set parameter i
set parameters
t function *f to graph *gr or histogram *h
get parameter i
get error on parameter i

TRandom3:

TF1(const char* name, const char* formula, Double_t xmin = 0, Double_t xmax = 1)

"gaus" "expo" "pol0" ... "pol9" "landau"

TF1(const char* name, void* fcn, Double_t xmin, Double_t xmax, Int_t npar)
.Eval(Double_t x)
Double_t .Derivative(Double_t x)
Double_t .Integral(Double_t a, Double_t b)
Double_t .GetRandom()
.SetParameter(Int_t i, Double_t parvalue)
.SetParameters(const Double_t* params)
gr->Fit(TF1 *f) or h->Fit(TF1 *f)|
Double_t .GetParameter(Int_t i)
Double_t .GetParError(Int_t i)

the calss used to generate random sequences of high quality

initialize random generator with random seed
TRandom(0)
initialize random generator with seed
TRandom(UInt_t seed)
get actual seed
UInt_t .GetSeed()
uniform random number ]0,x1]
Double_t .Uniform(Double_t x1=1)
uniform random number ]x1,x2]
Double_t .Uniform(Double_t x1, Double_t x2)
random number from binomial distribution
Int_t .Binomial(Int_t ntot, Double_t prob)
random Poisson number
Int_t .Poisson(Double_t mean)
random number from exponential
Double_t .Exp(Double_t tau)
random number from Gaussian distribution
Double_t .Gaus(Double_t mean=0, Double_t sigma=1)
pre-initialised random generator
gRandom points to global instance of TRandom3

TCanvas:

conguring the graphics canvas

create canvas of size ww x wh
TCanvas(const char* name, const char* title, Int_t ww, Int_t wh)
subdivide into pads
.Divide(Int_t nx = 1, Int_t ny = 1, Float_t xmargin = 0.01, Float_t ymargin = 0.01, Int_t color = 0)
chage to subpad
.cd(Int_t subpadnumber = 0)
update canvas
.Update()
mark as modied to trigger re-draw
.Modified(Bool_t flag = 1)
draw canvas
.Draw(Option_t* option = "") and .DrawClone

63

C. Most Relevant Classes and their Methods

TLegend:

the legend in a plot. Fundamental for the understanding of the contents

create Legend
TLegend(Double_t x1,Double_t y1,Double_t x2,Double_t y2,const char* header, Option_t* option = brNDC)
add an entry
.AddEntry(TObject* obj, const char* label, Option_t* option = lpf)
add text entry
.AddEntry(const char* name, const char* label, Option_t* option = lpf)
draw
.Draw() and .DrawClone();

TLatex:

LaTEX formatting

create Text
draw

TLatex(Double_t x, Double_t y, const char* text)

.Draw() and .DrawClone();

TFile:

le I/O
create le

change direcotry to e
write histogram *h to le
close le at the end
read histogram *h from le *f

TNtuple:

TFile(const char* fname, Option_t* option = "", const char* ftitle = "", Int_t compress = 1)
options " NEW" "CREATE" "RECREATE" "READ"

.cd()
h1->Write()
.Close()

TH1F *h1=(TH1F*)f.Get(const char* histname)

variables in ntuples

create

ll
initialize from le
plot variables
e.g. plot variable xi
e.g. plot variable with cut on others
e.g. 2-dim plot of variables xi and xj
ll existing histogram from ntuple

TNtuple(const char* name, const char* title, const char* varlist)

format varlist: "x0:x2:...:xn" (n<15)
.Fill(Float_t x0,Float_t x1=0,Float_t x2=0, ... ,Float_t x14=0)
.ReadFile(const char* filename)
.Draw(const char* varexp, const char* selection)
.Draw("xi")
.Draw("xi","xj<3")
.Draw("xi:xj")
.Project(const char* hname, const char* varexp, const char* selection = "")

TStyle

TSystem

global pointers gStyle and gSystem as instances of classes

and
show statistics box
gStyle->SetOptStat(11...1)
show t parameters in statistics box
gStyle->SetOptFit(11...1)
suppress title boxes on graphs and histograms
gStyle->SetOptTitle(0)
for animations: add pause in milliseconds
gSystem->Sleep(UInt_t t)

TVirtualFitter:

Fitting
set default tter, e. g. name="Minuit"
TVirtualFitter::SetDefaultFitter("(const char* name = "")
create Fitter instance
TVirtualFitter::Fitter(0,Int_t maxpar=25);
dene a parameter
Int_t .SetParameter(Int_t ipar,const char* parname,Double_t value,Double_t verr,Double_t vlow,Double_t vhigh)
set function to be minimized
.SetFCN(void (*)(Int_t&, Double_t*,Double_t&f,Double_t*,Int_t) fcn)
x a parameter
.FixParameter(Int_t ipar)
release parameter
.ReleaseParameter(Int_t ipar)
get pointer to active tter instance
static TVirtualFitter* .GetFitter()
interaction with tter
Int_t .ExecuteCommand(const char* command, Double_t* args, Int_t nargs)
example: start t with MINUIT:
double arglist[2]={5000,0.01}; .ExecuteCommand("MINIMIZE",arglist,2)
example: error evaluation MINUIT / MINOS:
ExecuteCommand("MINOS",arglist,0)
get pointer to covariance matrix
Double_t* .GetCovarianceMatrix() const
interaction with MINUIT via global pointer gMinuit of class

TMinuit

set DeltaChi2 value for error determination

get coutour line as TGraph

64

gMinuit->SetErrorDef(float DeltaChi2)
(TGraph*)gMinuit->Contour(npoints, int par1, int par2)

Contents
1

Motivation and Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ROOT Basics

2.1

ROOT as calculator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2

ROOT as Function Plotter

2.3

Controlling ROOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.4

Plotting Measurements

2.5

Histograms in ROOT

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.6

Interactive ROOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.7

ROOT Beginners' FAQ

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ROOT type declarations for basic data types

2.7.2

Congure ROOT at start-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10

2.7.3

ROOT command history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

2.7.4

ROOT Global Variables

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

ROOT Macros

. . . . . . . . . . . . . . . . . . . . . . . . .

10

2.7.1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10

3.1

General Remarks on ROOT macros

3.2

A more complete example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14

3.3

Summary of Visual eects

16

3.4

Welcome to ROOT

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

3.3.1

Colours and Graph Markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16

3.3.2

Arrows and Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16

3.3.3

Text

17

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Interpretation and Compilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

3.4.1

Compile a Macro with ACLiC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

3.4.2

Compile a Macro with g++

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

4.1

Read Graph Points from File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

4.2

Polar Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

4.3

2D Graphs

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

Histograms

5.1

Your First Histogram

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

5.2

Add and Divide Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24

5.3

Two-dimensional Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

File I/O

6.1

Storing ROOT Objects

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29

6.2

N-tuples in ROOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

6.2.1

Storing simple N-tuples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

6.2.2

Reading N-tuples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

6.2.3

Storing Arbitrary N-tuples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

6.2.4

Processing N-tuples Spanning over Several Files . . . . . . . . . . . . . . . . . . . . . . . .

33

6.2.5

. . . . . . . . . . . . . . . .

33

6.2.6

For the advanced user: Processing trees with a selector script

For power-users: Multi-core processing with PROOF lite . . .

6.2.7

Optimisation Regarding N-tuples

Functions and Parameter Estimation

. . . . . . . . . . . . . . . .

36

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

7.1

Fitting Functions to Pseudo Data

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7.2

Toy Monte Carlo Experiments

39

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

7.3

Fitting in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

65

Contents

A.2

66

47
47

8.1.1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

49

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

RooFiLab

A.1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PyROOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ROOT in Python

8.1

More

Python- less ROOT

Root-based tool for tting: RooFiLab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

A.1.1

Installation

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

A.1.2

Usage of

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

56

Examples with

RooFiLab
RooFiLab .

A.2.1

Straight-line t with correlated erros in x and y . . . . . . . . . . . . . . . . . . . . . . . .

57

A.2.2

Averaging correlated measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

58

A.2.3

Fit of a polynomyal to data with Poisson errors . . . . . . . . . . . . . . . . . . . . . . . .

59

A.2.4

Correlated measurements with full covariance matrix . . . . . . . . . . . . . . . . . . . . .

60

Markers, Colours, Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

B.1

Colour Wheel and Graph Markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61

B.2

Lines and Arrows

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

B.3

Latex Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

Most Relevant Classes and their Methods

BIBLIOGRAPHY

[1] ReneBrun and Fons Rademakers,

ROOT - An Object Oriented Data Analysis Framework,

Proceedings AI-

HENP'96 Workshop, Lausanne, Sep. 1996, Nucl. Inst. and Meth. in Phys. Res. A 389 (1997) 81-86. See also

http://root.cern.ch.
[2]

http://root.cern.ch/drupal/content/users-guide

[3]

http://root.cern.ch/drupal/content/reference-guide

[4]

http://root.cern.ch/drupal/content/cint

[5]

http://root.cern.ch/drupal/category/package-context/pyroot

[6]

http://www.math.keio.ac.jp/~matumoto/emt.html

67