Rf 4 0 2_Datahandling

Data and categories: tools for manipulation of (un)binned datasets

Author: Clemens Lange, Wouter Verkerke (C++ version)
This notebook tutorial was automatically generated with ROOTBOOK-izer from the macro found in the ROOT repository on Monday, January 17, 2022 at 10:02 AM.

In [ ]:
from __future__ import print_function
import ROOT
import math

WVE Add reduction by range

Binned (RooDataHist) and unbinned datasets (RooDataSet) share many properties and inherit from a common abstract base class (RooAbsData), provides an interface for all operations that can be performed regardless of the data format

In [ ]:
x = ROOT.RooRealVar("x", "x", -10, 10)
y = ROOT.RooRealVar("y", "y", 0, 40)
c = ROOT.RooCategory("c", "c")
c.defineType("Plus", +1)
c.defineType("Minus", -1)

Basic operations on unbinned datasetss

ROOT.RooDataSet is an unbinned dataset (a collection of points in N-dimensional space)

In [ ]:
d = ROOT.RooDataSet("d", "d", {x, y, c})

Unlike ROOT.RooAbsArgs (ROOT.RooAbsPdf, ROOT.RooFormulaVar,....) datasets are not attached to the variables they are constructed from. Instead they are attached to an internal clone of the supplied set of arguments

Fill d with dummy values

In [ ]:
for i in range(1000):
    x.setVal(i / 50 - 10)
    y.setVal(math.sqrt(1.0 * i))
    if i % 2:

    # We must explicitly refer to x,y, here to pass the values because
    # d is not linked to them (as explained above)
    if i < 3:
        print(x, y, c)
    d.add({x, y, c})


The get() function returns a pointer to the internal copy of the RooArgSet(x,y,c) supplied in the constructor

In [ ]:
row = d.get()

Get with an argument loads a specific data point in row and returns a pointer to row argset. get() always returns the same pointer, unless an invalid row number is specified. In that case a null ptr is returned

In [ ]:

Reducing, appending and merging

The reduce() function returns a dataset which is a subset of the original

In [ ]:
print("\n >> d1 has only columns x,c")
d1 = d.reduce({x, c})

print("\n >> d2 has only column y")
d2 = d.reduce({y})

print("\n >> d3 has only the points with y>5.17")
d3 = d.reduce("y>5.17")

print("\n >> d4 has only columns x, for data points with y>5.17")
d4 = d.reduce({x, c}, "y>5.17")

The merge() function adds two data set column-wise

In [ ]:
print("\n >> merge d2(y) with d1(x,c) to form d1(x,c,y)")

The append() function addes two datasets row-wise

In [ ]:
print("\n >> append data points of d3 to d1")

Operations on binned datasets

A binned dataset can be constructed empty, an unbinned dataset, or from a ROOT native histogram (TH1,2,3)

In [ ]:
print(">> construct dh (binned) from d(unbinned) but only take the x and y dimensions, ")
print(">> the category 'c' will be projected in the filling process")

The binning of real variables (like x,y) is done using their fit range 'get/setRange()' and number of specified fit bins 'get/setBins()'. Category dimensions of binned datasets get one bin per defined category state

In [ ]:
dh = ROOT.RooDataHist("dh", "binned version of d", {x, y}, d)

yframe = y.frame(Bins=10, Title="Operations on binned datasets")
dh.plotOn(yframe)  # plot projection of 2D binned data on y

Examine the statistics of a binned dataset

In [ ]:
print(">> number of bins in dh   : ", dh.numEntries())
print(">> sum of weights in dh   : ", dh.sum(False))

accounts for bin volume

In [ ]:
print(">> integral over histogram: ", dh.sum(True))

Locate a bin from a set of coordinates and retrieve its properties

In [ ]:
print(">> retrieving the properties of the bin enclosing coordinate (x,y) = (0.3,20.5) bin center:")

load bin center coordinates in internal buffer

In [ ]:
dh.get({x, y}).Print("v")
print(" weight = ", dh.weight())  # return weight of last loaded coordinates

Reduce the 2-dimensional binned dataset to a 1-dimensional binned dataset

All reduce() methods are interfaced in RooAbsData. All reduction techniques demonstrated on unbinned datasets can be applied to binned datasets as well.

In [ ]:
print(">> Creating 1-dimensional projection on y of dh for bins with x>0")
dh2 = dh.reduce({y}, "x>0")

Add dh2 to yframe and redraw

In [ ]:
dh2.plotOn(yframe, LineColor="r", MarkerColor="r")

Saving and loading from file

Datasets can be persisted with ROOT I/O

In [ ]:
print("\n >> Persisting d via ROOT I/O")
f = ROOT.TFile("rf402_datahandling.root", "RECREATE")

To read back in future session:

ROOT.TFile f("rf402_datahandling.root") d = (ROOT.RooDataSet*) f.FindObject("d")

In [ ]:
c = ROOT.TCanvas("rf402_datahandling", "rf402_datahandling", 600, 600)


Draw all canvases

In [ ]:
from ROOT import gROOT