Skip to content

R Version Control

Learning Objectives

  • Learn to use here for consistent file paths
  • Learn to use renv for reproducible environments
  • Develop best practices for R project management

Why Package Management?

Package management in R is crucial for:

  1. Reproducibility: Ensure same package versions across different machines
  2. Dependency Tracking: Keep track of all required packages
  3. Project Isolation: Avoid conflicts between different projects
  4. Collaboration: Make it easy for others to recreate your environment

No Absolute Paths with here

The here package solves the problem of absolute paths by providing a consistent way to reference files relative to your project root.

Setting Up here

  1. Install the package: 

    install.packages("here")

  2. In each R script or Rmd file, start with: 

    here::i_am('path/to/this/file')

Using here

Instead of absolute paths or relative paths that might break, use here::here():

# Bad practice
data <- read.csv("/Users/me/Projects/my_project/data/my_data.csv")

# Good practice
here::i_am('R/analysis.R')
data <- read.csv(here::here('data', 'my_data.csv'))

Example Script Structure

# In R/01_fit_models.R
here::i_am('R/01_fit_models.R')

# Load data
data <- read.csv(here::here('data', 'cleaned_data.csv'))

# Save results
save(results, file = here::here('output', 'model_results.RData'))

Version Control with renv

The renv package manages project-specific package dependencies, ensuring reproducibility across different environments.

Initial Setup

  1. Install renv: 

    install.packages("renv")

  2. Initialize in your project: 

    renv::init()

This creates: - renv.lock: Records package versions - .Rprofile: Activates renv - renv/: Contains project library

Key renv Commands

# Install packages recorded in renv.lock
renv::restore()

# Update renv.lock with current packages
renv::snapshot()

# Remove a package
renv::remove("package_name")

# Check status
renv::status()

Collaborative Workflow with renv

  1. Initial Setup (User A): 

    renv::init()
# Commit renv.lock, .Rprofile, and renv/activate.R

  2. New Collaborator (User B): 

    # After cloning repository
renv::restore()

  3. Adding New Packages (Any User): 

    install.packages("new_package")
renv::snapshot()
# Commit updated renv.lock

  4. Syncing Changes (All Users): 

    # After pulling updates
renv::restore()

  1. Project Setup: 

    # Create new RStudio project
# Initialize renv and here
install.packages(c("renv", "here"))
renv::init()
library(here)

# Document dependencies in DESCRIPTION file
usethis::use_description()

# Set up version control
usethis::use_git()

# Create standard directory structure
dir.create(here("data", "raw"), recursive = TRUE)
dir.create(here("data", "processed"), recursive = TRUE)
dir.create(here("R"))
dir.create(here("results", "figures"), recursive = TRUE)

  2. Development Workflow: 

    # Start new analysis script
library(here)
library(tidyverse)

# Use here for all file paths
data <- read_csv(here("data", "raw", "input.csv"))

# Install new package if needed
renv::install("newpackage")

# Work on your analysis
processed_data <- clean_data(data)
write_csv(processed_data, here("data", "processed", "clean.csv"))

# Save results using here
ggsave(here("results", "figures", "analysis_plot.pdf"))

# Update lock file
renv::snapshot()

# Document dependencies in DESCRIPTION
usethis::use_package("newpackage")

  3. Collaboration Workflow: 

    # Clone project
# Open project in RStudio (important for here to work!)

# Restore environment
renv::restore()

# Load packages
library(here)

# Work on project using relative paths
source(here("R", "analysis.R"))

# Update dependencies
renv::snapshot()

  4. Deployment/Sharing: 

    # Ensure all paths use here()
# Check no absolute paths remain:
grep -r "setwd\\|read\\.csv(" R/
grep -r "^[A-Za-z]:" R/  # Windows paths
grep -r "^/" R/          # Unix paths

# Ensure all dependencies are documented
renv::snapshot()

# Clean up unused packages
renv::clean()

# Bundle project for sharing
renv::bundle()

Advanced Tips

1. Using Multiple R Versions

Use .Rversion file: 

4.1.2

2. Custom Package Sources

In renv.lock: 

{
  "R": {
    "Version": "4.1.2",
    "Repositories": [
      {
        "Name": "CRAN",
        "URL": "https://cloud.r-project.org"
      },
      {
        "Name": "Custom",
        "URL": "https://mycran.org"
      }
    ]
  }
}

3. Continuous Integration

Example GitHub Actions workflow: 

on: [push, pull_request]

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      - uses: r-lib/actions/setup-r@v2
      - uses: r-lib/actions/setup-renv@v2
      - run: renv::restore()
      - run: Rscript -e "source('analysis/analysis.R')"

Resources