Developing R Packages from Scratch - GeeksforGeeks (2024)

R is considered as a powerful language for statistical computing, machine learning and and data analysis. Its major strength lies in the way that it allows its users to develop their own packages. Creating an R package from scratch can seem as a hectic task, but with a structured approach, it becomes interesting to create and at the same time, it becomes more manageble. This guide helps to developing an R package, from setting up the project structure to sharing your package with the R community.

Why Develop an R Package?

The reasons of creating R package can be categorized into four categories such as –

• Reuse: Package your functions and datasets for use across multiple projects.
• Distribution: Share your code with colleagues or the broader R community.
• Organization: Keep your code modular and well-organized.
• Version Control: Track changes and maintain different versions of your package

Now we implement the steps involved in creating a R package,from Scratch in R Programming Language.

Step 1: Load Necessary Libraries and Dataset

First, ensure you have the devtools and roxygen packages installed and loaded.

install.packages("devtools")library("devtools")devtools::install_github("klutometis/roxygen")library(roxygen2)

Step 2: Create a New Package

• Open RStudio.
• Go to File > New Project > New Directory > R Package.
• Enter the package name that you desire to give to your package.
• Choose the directory on where the package is to be created.
• Ensure that all the informations are correct , and then click create project.

After clicking the create project, you can now take a look at the final structure of your created package, including its components, which will look similar to the image given below.

• DESCRIPTION: Metadata about your package (e.g., title, description, author, version, etc).
• NAMESPACE: Controls which functions are exported from your package.
• R/: Directory where your R scripts (functions) are stored.
• man/: Directory for the documentation files, which will be automatically generated while creating the package.

Step 3: Write the Functions

• In RStudio, navigate to the R/ directory within your package project that you have created. Right-click on the directory and select “New File” > “R Script”.

• Begin coding your function in the newly created script. For example, let’s say you want to create a function named my_sum that calculates the sum of 2 numbers:

#' My First Function#'#' This function adds two numbers.#' #' @param x A numeric value.#' @param y A numeric value.#' @return The sum of x and y.#' @exportmy_function <- function(x, y) { return(x + y)}

Step 4: Document Your Functions

Documentation is important for any R package that is created as it ensures that users understand how to use your functions inside that package.Use the devtools package to generate the documentation and run the following command in the console.

devtools::document()

Output:

• The documentation files will be generated in the man/ directory of your package. Additionally, the NAMESPACE file will be updated to export the documented functions.

• The DESCRIPTION file should also be updated to reflect the package’s metadata, including the title, description, version, author, and dependencies. You can also verify the changes by visiting the description file in the package structure.

Step 5: Add tests

First Install the required package to ensure that the functions work as expected.

install.package("testthat")

Next Create a tests/ directory in the package.Write test cases in scripts saved in the tests/testthat/ directory.

Now write the test code in the test script.

#test-your_function.R test_that("add_two_numbers works correctly", { # Define the inputs and expected outputs input1 <- 2 input2 <- 3 expected_output1 <- 5 # Test if the function returns the correct sum expect_equal(add_two_numbers(input1, input2), expected_output1) # Test if the function throws an error for incorrect input wrong_input <- "a" expect_error(add_two_numbers(wrong_input, input2))})

Output:

 devtools::test()
ℹ Testing Mypackage
Error in install.packages : Updating loaded packages
✔ | F W S OK | Context
✔ | 2 | your_function 
══ Results ════════════════════════════════════════════
[ FAIL 0 | WARN 0 | SKIP 0 | PASS 2 ]

Step 6:Building and Checking the Package

Once your functions for the package and the documentation are in hand, it’s time to build and check the package for further usage and publications.

Build the Package

Use devtools::build() to create a .tar.gz file of your package, which can be shared or installed.

devtools::build()

Output:

devtools::build()
── R CMD build ────────────────────────────────────────
✔ checking for file 'C:\Users\srisa\OneDrive\Desktop\Rpackage\Mypackage/DESCRIPTION' (527ms)
─ preparing 'Mypackage': (604ms)
✔ checking DESCRIPTION meta-information ...
─ checking for LF line-endings in source and make files and shell scripts (537ms)
─ checking for empty or unneeded directories
 Omitted 'LazyData' from DESCRIPTION
─ building 'Mypackage_0.1.0.tar.gz'

Check the Package

Run devtools::check() command to ensure there are no errors or warnings. This command checks the package for the common issues and verifies that everything is working correctly.

devtools::check()

Output:

0 error ✔ | 0 warnings ✔ | 0 notes ✔

Step 7:Sharing the Package

After you have created the package sucessfully, you can share them with the R dev community out there who are waiting for many new packages that would reduce their work in several ways such as:

• Locally: Use install.packages(“path_to_your_package”) to install the package locally.

• GitHub: Push your package to a GitHub repository and use devtools::install_github(“username/repository”) to install it.

• CRAN: Submit your package to CRAN, the official R package repository, for wider distribution.

Conclusion

Developing an R package from scratch is a good practice that enhances your ability to organize, share, and reuse code. By following this article, you can create a well-documented, tested, and shareable R package. Whether you’re building tools for personal use or contributing to the R community, package development is an essential skill for any programmer who has R in his skill bucket.