(snapshot)=
# Exercise 9: Writing the n-tuple
**(10 minutes)**

If you've made changes to your n-tuple (defining new columns, applying
filters) and you want to write the revised n-tuple to a file, use the
`Snapshot()` method. 

**Here's an example**

:::{code-block} c++
:name: snapshot-c-example
:caption: An example using the `Snapshot` method (C++)

// Open a dataframe from an n-tuple within a file.
auto dataframe = ROOT::RDataFrame("tree1","experiment.root");

// Make a change to the n-tuple.
auto pzcut = dataframe.Filter("pz < 145");

// Write the revised n-tuple with the name "pzcut" to file
// "analysis.root".
pzcut.Snapshot("pzcut","analysis.root");
:::

**Here's the Exercise**

In the past few pages, we've built a collection of new variables
($p_{T}$, $\theta$, $E_{meas}$, $E_{loss}$) and useful analysis cuts
($\chi^{2} < 1.5$, $\theta < 0.15$). Derive an n-tuple with all these
new columns and cuts, and write it as n-tuple `analyzed` to file
{file}`analysis.root`.

:::{tip}
If you look up `Snapshot()` on the [RDataFrame web page](https://root.cern/doc/master/classROOT_1_1RDataFrame.html)
you'll see that it's classified as an {dfn}`instant action`. This means
that the event loop will be executed when the `Snapshot()` is executed.

If, for some reason, you want a lazy evaluation of `Snapshot()`, click on the highlighted word
`Snapshot()` on that page. It will take you to a different page; search for `fLazy` on that new page.
:::

:::{figure-md} effect_size-fig
:align: center

<img src="https://imgs.xkcd.com/comics/effect_size.png" alt="xkcd effect_size" width="50%">

<https://xkcd.com/2755/> by Randall Munroe
:::