R-Environment-Notes

Notes on using renv to manage virtual environments in R.

This seems much more straightforward than python so it's all below in the README.

---

Check if renv is installed (if you've never installed it it won't be)

renv is not part of the base R installation. It is a contributed package available from CRAN (the Comprehensive R Archive Network), just like dplyr or ggplot2. Check whether renv is installed by trying to load it with the library() function.

library(renv)

What the results mean:

• If it's INSTALLED: The command will run silently without any output, or it might print a startup message from the package. You won't get an error.

• If it's NOT INSTALLED: You will get a clear error message.

  Error in library(renv) : there is no package called ‘renv’
  

If its not installed install it

You only need to do this once on your computer. You install it into your main, global R library like any other package.

Open your R or RStudio console and run:

install.packages("renv")

Once this command is finished, the renv package is available for you to use in any of your R projects.

Step 1: Initialize renv in a Project

You only need to do this once when you start a new project or decide to use renv with an existing one.

• Navigate to your project's root directory in the R console.
• Run the init() command.

# Run this once per project
renv::init()

This command will:

1. Create a project-specific library inside a new renv/ folder.
2. Create or modify the project's .Rprofile file to ensure renv is loaded automatically whenever you open the project.
3. Discover the packages you are currently using in your scripts, install them into the new library, and create an initial renv.lock file by taking a snapshot.

Note that renv might install way more packages than you were expecting it to - that is because it installs all the package dependencies too.

Step 2: Install or Remove Packages

As you work on your project, you'll install new packages. You can do this in the standard way. renv is smart enough to intercept this.

# Install a new package
install.packages("dplyr")

# Or, install multiple packages
install.packages(c("ggplot2", "stringr"))

# To remove a package
remove.packages("some_old_package")

These packages will be installed into your project's private renv library, not your global R library.

Step 3: Save Your Changes (Take a Snapshot)

After you have installed, updated, or removed packages and are happy with the state of your project, you must save these changes to the lockfile. This is the crucial step for reproducibility.

# Save the current state of your project library to renv.lock
renv::snapshot()

When you run this, renv will:

• Scan your R scripts for required packages.
• Compare those packages with the state of your project library.
• Show you which packages will be added, removed, or changed in renv.lock.
• Ask for your confirmation before writing to the lockfile.

Best Practice: Think of renv::snapshot() like git commit. You do it whenever you've reached a stable, meaningful state in your project's dependencies. You should add renv.lock to your Git repository.

Step 4: Restore the Environment on a New Machine

This is the "payoff" step. Imagine a collaborator (or your future self on a new computer) has just cloned your project from Git. The repository contains your R scripts and the renv.lock file.

To get the exact same package environment, they just need to:

1. Open the project in R or RStudio.
2. Run the restore() command. (If renv isn't activated it it will prompt you to activate renv first).

# Read renv.lock and install packages to match it
renv::restore()

renv will read renv.lock, see which packages are needed (and their exact versions), and install them. Your collaborator now has a perfect replica of your package environment, avoiding "it works on my machine" problems.

Other Useful Commands

• renv::status(): Shows you the difference between the packages defined in your renv.lock and the packages currently installed in your library. It's great for checking if you need to run snapshot() or restore().
• renv::history(): Shows the history of changes to the lockfile.
• renv::revert(): Reverts changes to the lockfile, similar to git revert.