Chapter 3 R-Scripts

As your analysis conducted in R gets more complicated (at least in the number of expressions to evaluate), you may want to save your work as an R script, rather than typing one command after the other into the console. An R Script is really just a plain text file containing R code that can be executed. A major benefit of R scripts is that you preserve the analysis you did and can easily re-run it again - that is not only going to save you time, but it will also make your work reproducible.

RStudio is fantastic for working with R scripts. Just click on File -> New File -> R Script in the menu and a new file will be opened in the source pane. There, you can type your R code and save it for later use, but also send it to the console for execution.

For instance, copy the following code into an empty R script:

x <- 10
y <- 20
z <- x + y

You can now execute this R-script in one of two ways in RStudio:

You can send line by line to the console. for this, place the cursor anywhere on the chosen line an click the “Run” button (or press Ctrl + Enter as a shortcut). The cursor will automatically jump to the next line, so you can just keep pressing the “Run” button (or pressing Ctrl + Enter) until you reach the end of the script.
Alternatively, you can ask for the whole document to be executed. For this, simply press the “Source” button (or press Ctrl+ Shift+ Enter as a shortcut).

Cool, no?

Comments

As your R code will grow, it will become very important to comment it properly. Indeed, a variable day_1 will maybe mean a a lot to you while writing your code, but will you remember its meaning when going back to it after one year? Similarly, you maybe side with a particular statistical analysis, but will you remember the reasoning a year later?

Most peopel don’t - which is why most people should add comments to their code. In R, comments always start with a # symbol. This tells R that everything after the # is not R code and is to be ignored.

> x <- 7
> #x <- 5
> x
[1] 7

You can add comments at the end of a line too

> num_replicates <- 10 # I chose 10 to balance statistical power with sampling effort

Since R scripts can get very long, you can also use comments to structure them or leave other worthy notes. Here is an example:

#------------------------------------------------
# My first R script
#
# This script illustrates what I learned so far
#------------------------------------------------

# Assigning variables
num_replicates <- 4 
b <- 8 # TODO: This is a bad variable name - think of a better one

# Calling functions
z <- log(a) - log(b);
print(exp(z)) # we expect this to be 0.5
[1] 0.5

3.0.1 Exercises: R-Scripts

See Section 18.0.4 for solutions.

Create an R-script that assigns the values $π = 3.1415926 \dots$ , 10 and $e = 2.718282 \dots$ to variables my_data, base_1 and base_2 and then prints the log of my_data with both bases. Use comments to explain your script and save it as myFirstRScript.r on your computer.
Execute your script myFirstRScript.r both line-by-line as well as in one go (i.e. source it).