03-tfile-read-write-ttree.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# The ROOT file"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "* With ROOT, objects can be written to files\n",
    "\n",
    "* ROOT provides its own file class, [TFile](https://root.cern/doc/master/classTFile.html), to interact with these files\n",
    "\n",
    "* ROOT files are _binary_ and can be transparently _compressed_ to reduce disk usage\n",
    "\n",
    "* ROOT files have a logical “file-system-like” structure\n",
    "\n",
    "  * E.g. a directory hierarchy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's start with importing ROOT as usual"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "import ROOT"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "\n",
    "This is how you create a `TFile`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "f = ROOT.TFile(\"my_file.root\", \"RECREATE\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<center><img src=\"images/tfile1.png\"><center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "and how you close it (note that when `f` is destroyed, the file is closed automatically):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "f.Close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Note that you can use the Python context manager syntax for opening a TFile, which will automatically run the necessary cleanup for you:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "with ROOT.TFile(\"my_file.root\", \"RECREATE\") as myfile:\n",
    "    # do something with the file inside the scope\n",
    "    pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "With the following example we demonstrate how to write an object inside a file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "with ROOT.TFile(\"my_file.root\", \"RECREATE\") as f:\n",
    "    h = ROOT.TH1D(\"my_histo\", \"Example histogram\", 100, -4, 4)\n",
    "    f.WriteObject(h, h.GetName())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "\n",
    "The `\"my_histo\"` argument of the `TH1D` constructor is the name of the histogram, and it is also how it will be identified inside the file, we'll see that in a minute."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "We should now have a file called `my_file.root` in the current directory. We will check that by using the `%%bash` magic, which allows us to run bash commands from a cell:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%bash\n",
    "ls -l my_file.root"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "We can also use the `rootls` command to inspect the contents of the ROOT file. See how the file contains an object called `my_histo` of type `TH1D`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%bash\n",
    "rootls -l my_file.root"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Finally, let's see how we can programmatically retrieve the histogram we just wrote in the file. \n",
    "\n",
    "We can access the histogram by its name using `TFile::Get()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "with ROOT.TFile(\"my_file.root\") as f: # READ is the default mode\n",
    "    h = f.Get(\"my_histo\")\n",
    "    print(h)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# The HEP dataset\n",
    "\n",
    "High Energy Physics data is made of many statistically independent collision events. \n",
    "\n",
    "Laying data into an \"event class\", then serialise and write out `N` instances of the class into a file would be very inefficient. \n",
    "\n",
    "In ROOT, a dataset is organised columns that can store elements of any C++ type:\n",
    "* fundamental types: `int`, `float`\n",
    "* C++ standard collections: `std::vector`, `std::map`\n",
    "* User created C++ classes\n",
    "\n",
    "The ROOT dataset is represented by the `TTree` class and is often simply called a tree. Columns in the dataset are instances of the `TBranch` class (often referred to as \"branches\")."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "<center><img src=\"images/dataset.png\"></center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "- A `TTree` dataset can be written to a `TFile` (just like any other C++ object). \n",
    "\n",
    "- The ROOT format is logically and physically (on disk) a columnar format. \n",
    "\n",
    "- Different columns can be read from disk independently. \n",
    "\n",
    "- This translates into faster IO performance with respect to other dataset formats (HDF5, SQL)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "%%bash\n",
    "rootls -l data/example_file.root"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}