Run Multiple R Versions Side-by-Side (and Switch Without Breaking Anything)

You can install multiple R versions on the same computer and switch between them. This lets you use the latest R for new projects while keeping an older version for existing work that depends on it.

A new R version comes out every April. You want the new features, but updating might break packages your current project relies on. The solution: install both versions and switch as needed.

Why You'd Need Multiple R Versions

cat("Common scenarios:\n\n") cat("1. UPGRADE SAFETY\n") cat(" New R version released, but your thesis depends on R 4.3\n") cat(" → Keep 4.3 for the thesis, install 4.4 for new work\n\n") cat("2. PACKAGE COMPATIBILITY\n") cat(" A critical package hasn't been updated for the latest R\n") cat(" → Run that package on the old R, everything else on new\n\n") cat("3. REPRODUCIBILITY\n") cat(" Reviewer asks you to verify results with the exact R version\n") cat(" → Install that specific version and re-run\n\n") cat("4. TEAM ALIGNMENT\n") cat(" Your team uses R 4.3, you have R 4.4\n") cat(" → Match their version to avoid 'works on my machine' bugs\n")

How R Versions Coexist

On all operating systems, each R version installs to its own directory. They don't overwrite each other:

cat("=== Installation paths ===\n\n") cat("Windows:\n") cat(" C:/Program Files/R/R-4.3.3/\n") cat(" C:/Program Files/R/R-4.4.2/\n\n") cat("Mac:\n") cat(" /Library/Frameworks/R.framework/Versions/4.3-arm64/\n") cat(" /Library/Frameworks/R.framework/Versions/4.4-arm64/\n\n") cat("Linux (Ubuntu, with rig):\n") cat(" /opt/R/4.3.3/\n") cat(" /opt/R/4.4.2/\n\n") cat("Each version has its own:\n") cat(" - R executable\n") cat(" - Package library\n") cat(" - Configuration files\n")

Method 1: rig — The Best Tool (All Platforms)

rig (R Installation Manager) is the modern way to manage R versions. It works on Windows, Mac, and Linux:

# Install rig (one-time setup)

# Windows (PowerShell as admin):
# winget install posit.rig

# Mac (Homebrew):
# brew tap r-lib/rig
# brew install --cask rig

# Ubuntu/Debian:
# curl -Ls https://github.com/r-lib/rig/releases/download/latest/rig-linux-latest.tar.gz |
#   sudo tar xz -C /usr/local

# After installing rig:
rig list              # Show installed R versions
rig add 4.3           # Install R 4.3.x
rig add 4.4           # Install R 4.4.x
rig default 4.4       # Set default version
rig resolve 4.3       # Find exact version number

Switching versions with rig

cat("=== rig commands ===\n\n") cat("rig list\n") cat(" Shows all installed R versions\n") cat(" * marks the default\n\n") cat("rig default 4.3\n") cat(" Sets R 4.3 as the default for new terminals\n\n") cat("rig rstudio 4.3\n") cat(" Opens RStudio using R 4.3 (instead of default)\n") cat(" This is the killer feature — project-specific R versions!\n\n") cat("rig add release\n") cat(" Installs the latest stable R release\n\n") cat("rig add devel\n") cat(" Installs the development version of R\n")

Method 2: Windows — Side-by-Side Installation

On Windows, you can install multiple R versions without any special tools:

cat("=== Windows: Manual side-by-side ===\n\n") cat("Step 1: Install R 4.3.3\n") cat(" Download from CRAN, install to default path\n") cat(" → C:/Program Files/R/R-4.3.3/\n\n") cat("Step 2: Install R 4.4.2\n") cat(" Download from CRAN, install to default path\n") cat(" → C:/Program Files/R/R-4.4.2/\n") cat(" Both versions now exist in C:/Program Files/R/\n\n") cat("Step 3: Switch in RStudio\n") cat(" Tools → Global Options → General → R version → Change\n") cat(" Select the version you want → Restart RStudio\n\n") cat("The installr package can automate upgrades:\n") cat(" install.packages('installr')\n") cat(" installr::updateR() # Downloads and installs latest R\n")

Method 3: Mac — Framework Versions

Mac's R installer places each version in a separate framework directory:

cat("=== Mac: Side-by-side ===\n\n") cat("Step 1: Download both .pkg installers from CRAN\n") cat("Step 2: Install both — they go to separate directories\n\n") cat("Step 3: Switch in RStudio\n") cat(" RStudio → Preferences → General → R version → Change\n") cat(" Select the version you want → Restart RStudio\n\n") cat("Step 4: Switch on command line (optional)\n") cat(" # Create a symbolic link to switch the default:\n") cat(" sudo ln -sf /Library/Frameworks/R.framework/Versions/4.3-arm64 \\\n") cat(" /Library/Frameworks/R.framework/Versions/Current\n\n") cat("Recommended: Use rig instead — it handles all this automatically.\n")

Method 4: Linux — Multiple Installs with rig

On Linux, managing multiple R versions without rig requires manual compilation. With rig, it's simple:

cat("=== Linux: rig is essential ===\n\n") cat("# Install multiple versions\n") cat("sudo rig add 4.3\n") cat("sudo rig add 4.4\n\n") cat("# Each version gets its own directory:\n") cat("# /opt/R/4.3.3/bin/R\n") cat("# /opt/R/4.4.2/bin/R\n\n") cat("# Switch default\n") cat("sudo rig default 4.4\n\n") cat("# Run a specific version directly\n") cat("R-4.3 # Runs R 4.3\n") cat("R-4.4 # Runs R 4.4\n")

Managing Package Libraries

Each R version has its own package library. When you switch versions, packages from one version aren't automatically available in the other:

# Check your current R version and library path cat("R version:", R.version.string, "\n") cat("Library path:", .libPaths(), "\n") # See installed packages pkgs <- installed.packages() cat("Installed packages:", nrow(pkgs), "\n") cat("First 10:", head(rownames(pkgs), 10), "\n")

Reinstalling packages after upgrading

cat("=== Package migration strategies ===\n\n") cat("Strategy 1: Reinstall everything (cleanest)\n") cat(" # On the OLD version, save package list:\n") cat(" old_pkgs <- installed.packages()[, 'Package']\n") cat(" saveRDS(old_pkgs, 'my_packages.rds')\n\n") cat(" # On the NEW version, reinstall:\n") cat(" old_pkgs <- readRDS('my_packages.rds')\n") cat(" install.packages(old_pkgs)\n\n") cat("Strategy 2: Shared library (risky)\n") cat(" # Point both versions to the same library\n") cat(" # .libPaths('/path/to/shared/library')\n") cat(" # Warning: can cause compatibility issues\n\n") cat("Strategy 3: renv per project (recommended)\n") cat(" # Each project has its own package library\n") cat(" # renv::init() in each project\n") cat(" # Completely isolated — version switches are safe\n")

renv: The Best Solution for Projects

cat("=== renv: Project-level package management ===\n\n") cat("# Initialize renv in a project:\n") cat("# renv::init()\n\n") cat("# renv creates a project-local library:\n") cat("# project/renv/library/R-4.3/...\n\n") cat("# Snapshot your packages:\n") cat("# renv::snapshot() → saves to renv.lock\n\n") cat("# Restore on another machine (or after R upgrade):\n") cat("# renv::restore() → installs exact package versions from renv.lock\n\n") cat("Benefits:\n") cat(" - Each project is isolated\n") cat(" - Switching R versions doesn't affect projects\n") cat(" - renv.lock is a reproducible recipe\n") cat(" - Collaborators get the exact same packages\n")

Checking Your R Version

# Multiple ways to check which R you're running cat("R.version.string:", R.version.string, "\n") cat("R.version$major:", R.version$major, "\n") cat("R.version$minor:", R.version$minor, "\n") # Check if a minimum version is met required <- "4.1.0" current <- paste(R.version$major, R.version$minor, sep = ".") cat("\nCurrent version:", current, "\n") cat("Required:", required, "\n") cat("Meets requirement:", compareVersion(current, required) >= 0, "\n")

# Useful for scripts that need a specific version check_r_version <- function(min_version = "4.1.0") { current <- paste(R.version$major, R.version$minor, sep = ".") if (compareVersion(current, min_version) < 0) { stop(paste("This script requires R >=", min_version, "but you have R", current)) } cat("R version check passed:", current, ">=", min_version, "\n") } check_r_version("4.0.0")

When to Upgrade R

cat("=== R release cycle ===\n\n") cat("Major releases (4.x.0): Every April\n") cat(" - May contain breaking changes\n") cat(" - Packages may need updates\n") cat(" - Wait 2-4 weeks for package ecosystem to catch up\n\n") cat("Patch releases (4.x.y): Throughout the year\n") cat(" - Bug fixes only\n") cat(" - Safe to update immediately\n") cat(" - Packages remain compatible\n\n") cat("Recommendation:\n") cat(" 1. Update patch releases immediately\n") cat(" 2. Wait 2-4 weeks after major releases\n") cat(" 3. Keep the old version until your key packages work on the new one\n") cat(" 4. Use renv for projects that must be reproducible\n")

Practice Exercises

Exercise 1: Version Check Script

# Exercise: Write a function that: # 1. Prints the current R version # 2. Checks if the native pipe |> is available (requires R >= 4.1) # 3. Checks if the lambda syntax \(x) is available (requires R >= 4.1) # 4. Reports which features are available # Write your code below:

Click to reveal solution

# Solution check_features <- function() { ver <- paste(R.version$major, R.version$minor, sep = ".") cat("R version:", ver, "\n\n") # Native pipe |> (R 4.1+) has_pipe <- compareVersion(ver, "4.1.0") >= 0 cat("Native pipe |>:", if (has_pipe) "Available" else "Not available", "\n") # Lambda \(x) syntax (R 4.1+) has_lambda <- compareVersion(ver, "4.1.0") >= 0 cat("Lambda \\(x):", if (has_lambda) "Available" else "Not available", "\n") # Underscore placeholder _ (R 4.2+) has_placeholder <- compareVersion(ver, "4.2.0") >= 0 cat("Pipe placeholder _:", if (has_placeholder) "Available" else "Not available", "\n") # Raw string literals r"(...)" (R 4.0+) has_raw_string <- compareVersion(ver, "4.0.0") >= 0 cat("Raw strings r\"(...)\":", if (has_raw_string) "Available" else "Not available", "\n") } check_features()

Explanation: compareVersion() compares version strings correctly (it knows "4.10" > "4.9"). Each R feature was introduced in a specific version — this function checks which ones are available in your current installation.

Exercise 2: Package Inventory

# Exercise: Write a script that creates a "package backup" — a data frame # listing all installed packages with their versions and source (CRAN/other). # This is what you'd save before upgrading R. # Write your code below:

Click to reveal solution

# Solution pkg_info <- installed.packages() backup <- data.frame( package = pkg_info[, "Package"], version = pkg_info[, "Version"], priority = pkg_info[, "Priority"], stringsAsFactors = FALSE ) # Separate base/recommended from user-installed backup$type <- ifelse( is.na(backup$priority), "user-installed", backup$priority ) cat("Package inventory:\n") cat(" Base packages:", sum(backup$type == "base"), "\n") cat(" Recommended:", sum(backup$type == "recommended"), "\n") cat(" User-installed:", sum(backup$type == "user-installed"), "\n") cat(" Total:", nrow(backup), "\n\n") # Show user-installed packages (these need reinstalling after upgrade) user_pkgs <- backup[backup$type == "user-installed", ] cat("User-installed packages (first 15):\n") print(head(user_pkgs[, c("package", "version")], 15)) # To save: saveRDS(user_pkgs$package, "my_packages.rds")

Explanation: installed.packages() returns a matrix with all package metadata. Base and recommended packages come with R — you only need to reinstall "user-installed" ones after upgrading. Saving this list before upgrading lets you restore your setup quickly.

Summary

Method	Platform	Best for
rig	All	Modern, recommended approach
Manual side-by-side	Windows	Simple, no extra tools
RStudio version selector	All	Quick switching without terminal
renv	All	Project-level package isolation

The recommended workflow:

Install rig for managing R versions
Use renv for project-level package isolation
Keep one version behind for critical projects
Wait 2-4 weeks after major releases before upgrading

FAQ

Do I need to reinstall all packages when I upgrade R?

For minor upgrades (4.3.2 → 4.3.3): no, packages are compatible. For major upgrades (4.3 → 4.4): usually yes, especially packages that contain compiled C/C++ code. The safest approach is to save your package list, upgrade, and reinstall.

Can RStudio use different R versions for different projects?

Not natively per-project, but with rig you can run rig rstudio 4.3 to open RStudio with a specific R version. Or use RStudio's global setting: Tools → Global Options → General → R version.

What's the difference between rig and renv?

rig manages R versions (the engine). renv manages packages (the libraries). They work together: rig switches R versions, renv ensures each project has the right packages for its R version.

Will my old scripts break on a new R version?

Usually no. R is very backwards-compatible. The most common breaking changes involve: (1) stringsAsFactors defaulting to FALSE in R 4.0, (2) new pipe |> syntax in R 4.1, (3) changes in default random number generation. Check the R NEWS file for specific version changes.

How do I check which R version a project was built with?

If the project uses renv, check renv.lock — it records the R version. Otherwise, look for sessionInfo() output in reports or README files.

What's Next?

Version management is part of the broader reproducibility toolkit:

Positron vs RStudio — the new IDE alternative
R Project Structure — organize your R projects properly
renv and Reproducibility — lock down package versions for reproducible research

r-statistics.co by Selva Prabhakaran