4 Working with a script

In this chapter, you will

learn why you should use scripts for R code
learn how to run code in a script
learn how and why to use comments in code
learn some tips for making readable code

4.1 What is a script?

You can code directly in the RStudio console. This is useful for throwaway code that you never want to run again. If you want to be able to run the code again (and you probably do) you should put your code in a script (or equivalently a code chunk in a quarto document).

Practically, a script is a plain text file where you write your code, whether it contains a handful of lines or dozens of them. It is an evolving document which not only helps you keep track of your code, but also your workflow.

With time, you will realize that a script is a lot of things at the same time:

it is a whiteboard where you try coding for something and correct mistakes whenever you find out things do not work as expected,
it contains your coding history, where all the steps from loading a data set to printing the final output are chronologically exposed,
it is the key file that you may share with collaborators, etc,
it is guarantee that your work is reproducible, meaning that you can run your code on your data set again and again, and obtain the same result, consistently.

A simple script may look like this:

# activate tidyverse
library(tidyverse)

# load the data from external file
Veronica_Vestland <- read_delim("Veronica_Vestland.csv", delim = ",")

# calculate the mean and standard deviation of Sepal.Length for each Location
mean_sd_SL <- Veronica_Vestland |>
           group_by(location) |>
           summarise(mean(Sepal.Length), sd(Sepal.Length))

# print the result
mean_sd_SL

# draw boxplot Sepal.Length for each Location
ggplot(Veronica_Vestland,
       aes(x = Location, y  = Sepal.Length, fill = Location)) +
  geom_boxplot()

You will be able to write similar code to this very soon.

4.2 Making a new script

You can make a new script with

4.3 Running the code

Writing code in a script does not do anything per se. To tell R to do something, you must place the cursor anywhere on a line of code and

press the Run button above the script
type to run that line

This will run the line of code and any syntactically connected to it. You can also select some code (perhaps a part of a line or several lines) and run it with either of the above methods.

The result of your command(s) will appear in the tab Console if the commands are intended to print something, and/or in the tab Plots if the commands generate a plot.

4.4 Code versus comment

In the script above, there are two types of lines: those that start with the symbol #, and those that do not.

Let’s start with the lines that do not start with a #. They are the real code, the commands that manipulate the data. Right now these lines do not mean much to you, but in fact, each of them commands R to “do something specific” with your data. That “something specific” is defined by functions which are followed by parentheses – function(). In the R language, functions are verbs in your sentences, the data are their subject. For example, in the code above, library(tidyverse) commands R to activate the package tidyverse found in the package library.

The lines that start with a # are comments. They do not code for anything at all. When you run the script via a console, R simply ignores them. So use comments to keep track of what you do with the code. Write what the point of each real code line is, what you plan to do. That way, you will always remember what you originally intended to code for.

The symbol # is also convenient to prevent R from running a specific code line or chunk, without having to delete that line. Indeed if you place a # in front of any line, the console will consider it as a comment, and simply skip it.

In the following example, each line was originally written to activate a different package:

library(ggplot2)
library(tidyr)
# library(tidyverse)
library(readr)

However, the third line has been commented with a #. Consequently, only the packages ggplot2, tidyr, and readr are activated; tidyverse will be ignored. If you want to comment out or uncomment many lines of code at once, you can use the Rstudio shortcut .

4.5 Writing readable code

Code can be difficult to read when you try working on it again after a few weeks or months.

Here are three tips for making it easier

use good comments
use good object names (Section 8.1.4)
codeishardtoreadifwithoutwhitespace

In the section First Steps in R, you will learn to write in the R language. We strongly advise you to work in scripts, and make extensive use of comments from the start. This is considered good coding practice, and will save you quite some time and energy.

4.6 Special characters

R uses some special characters, including ~, {, [, and $ (see Chapter 25 for a more complete list). These characters are usually easy to find on a Windows computer. If you are using a Mac, make sure you can find these.

Quiz

Question 1: Your R console shows ‘+’ instead of the usual ‘>’ prompt. What does this mean?

R is broken and need to be re-installedthe current command is unfinished. Perhaps a bracket or quote mark is missing.R is now a calculator and is waiting for a number

See Section 2.3.3.1. Usually best to press escape.

Contributors

Jonathan Soulé
Richard Telford

# Working with a script {#sec-working-script} ::: callout-note ## In this chapter, you will - learn why you should use scripts for R code - learn how to run code in a script - learn how and why to use comments in code - learn some tips for making readable code ::: ## What is a script? You can code directly in the RStudio console. This is useful for throwaway code that you never want to run again. If you want to be able to run the code again (and you probably do) you should put your code in a script (or equivalently a code chunk in a [quarto document](https://biostats-r.github.io/biostats/quarto/index.html)). Practically, a script is a plain text file where you write your code, whether it contains a handful of lines or dozens of them. It is an evolving document which not only helps you keep track of your code, but also your workflow. With time, you will realize that a script is a lot of things at the same time: - it is a whiteboard where you try coding for something and correct mistakes whenever you find out things do not work as expected, - it contains your coding history, where all the steps from loading a data set to printing the final output are chronologically exposed, - it is the key file that you may share with collaborators, etc, - it is guarantee that your work is reproducible, meaning that you can run your code on your data set again and again, and obtain the same result, consistently. A simple script may look like this: ```{r commented-script, echo=TRUE, eval=FALSE} # activate tidyverse library(tidyverse) # load the data from external file Veronica_Vestland <- read_delim("Veronica_Vestland.csv", delim = ",") # calculate the mean and standard deviation of Sepal.Length for each Location mean_sd_SL <- Veronica_Vestland |> group_by(location) |> summarise(mean(Sepal.Length), sd(Sepal.Length)) # print the result mean_sd_SL # draw boxplot Sepal.Length for each Location ggplot(Veronica_Vestland, aes(x = Location, y = Sepal.Length, fill = Location)) + geom_boxplot() ``` You will be able to write similar code to this very soon. ## Making a new script You can make a new script with <ol class="breadcrumb"> <li class="breadcrumb-item">File</li> <li class="breadcrumb-item">New File</li> <li class="breadcrumb-item">R Script</li> </ol> ## Running the code Writing code in a script does not do anything *per se*. To tell R to do something, you must place the cursor anywhere on a line of code and - press the `Run` button above the script - type {{< kbd mac=Command-Return win=Control-Enter linux=Ctrl-Enter >}} to run that line This will run the line of code and any syntactically connected to it. You can also select some code (perhaps a part of a line or several lines) and run it with either of the above methods. The result of your command(s) will appear in the tab `Console` if the commands are intended to print something, and/or in the tab `Plots` if the commands generate a plot. ## Code versus comment In the script above, there are two types of lines: those that start with the symbol `#`, and those that do not. Let's start with the lines that do not start with a `#`. They are the *real* code, the commands that manipulate the data. Right now these lines do not mean much to you, but in fact, each of them commands R to "do something specific" with your data. That "something specific" is defined by *functions* which are followed by parentheses -- `function()`. In the R language, functions are verbs in your sentences, the data are their subject. For example, in the code above, `library(tidyverse)` commands R to *activate* the package tidyverse found in the package library. The lines that start with a `#` are *comments*. They do not code for anything at all. When you run the script via a console, R simply ignores them. So use comments to keep track of what you do with the code. Write what the point of each *real* code line is, what you plan to do. That way, you will always remember what you originally intended to code for. The symbol `#` is also convenient to prevent R from running a specific code line or chunk, without having to delete that line. Indeed if you place a `#` in front of any line, the console will consider it as a comment, and simply skip it. In the following example, each line was originally written to activate a different package: ```{r commenting-to-deactivate, echo=TRUE, eval=FALSE} library(ggplot2) library(tidyr) # library(tidyverse) library(readr) ``` However, the third line has been *commented* with a `#`. Consequently, only the packages `ggplot2`, `tidyr`, and `readr` are activated; `tidyverse` will be ignored. If you want to comment out or uncomment many lines of code at once, you can use the Rstudio shortcut {{< kbd mac=Shift-Command-C win=Control-Shift-C linux=Ctrl-Shift-C >}}. ## Writing readable code Code can be difficult to read when you try working on it again after a few weeks or months. Here are three tips for making it easier - use good comments - use good object names (@sec-naming-objects) - codeishardtoreadifwithoutwhitespace In the section First Steps in R, you will learn to write in the R language. We strongly advise you to work in scripts, and make extensive use of comments from the start. This is considered good coding practice, and will save you quite some time and energy. ## Special characters R uses some special characters, including `~`, `{`, `[`, and `$` (see @sec-bestiary for a more complete list). These characters are usually easy to find on a Windows computer. If you are using a Mac, make sure you can find these. ```{r} #| label: setup-quiz #| file: !expr here::here("Templates/webex.R") #| include: false ``` ::: {.callout-note} ## Quiz ```{r} #| label: quiz #| results: asis #| echo: false questions <- list( list( question = "Your R console shows '+' instead of the usual '>' prompt. What does this mean?", choice = c( "R is broken and need to be re-installed", answer = "the current command is unfinished. Perhaps a bracket or quote mark is missing.", "R is now a calculator and is waiting for a number" ), hint = "See @sec-console-tab. Usually best to press {{< kbd escape >}}. " ) ) print_multichoice(questions) ``` ::: ::: {.column-margin} ### Contributors {.unlisted .unnumbered} - Jonathan Soulé - Richard Telford :::