🔬 Tidyverse Series – Post 9: Functional Programming in R with purrr Link to heading

🛠 Why {purrr}? Link to heading

Iteration in R can be clunky using for-loops, especially when working with lists and complex data structures. {purrr} provides elegant functional programming tools that make iteration cleaner, safer, and more expressive.

🔹 Why Use {purrr}? Link to heading

✔️ Replaces slow and verbose for-loops
✔️ Works seamlessly with lists and tibbles
✔️ Encourages a functional programming approach
✔️ Improves error handling in iterative tasks
✔️ Simplifies nested operations

Let’s explore the key functions that make {purrr} a game-changer.

📚 Key {purrr} Functions Link to heading

📊 Example 1: Applying a Function to a List of Values Link to heading

Imagine we need to take the square root of a list of numbers, handling potential errors.

➡️ Base R approach using a for-loop: Link to heading

numbers <- list(4, 9, 16, 25, -1)
results <- c()

for (x in numbers) {
  results <- c(results, sqrt(x))
}

✔️ Works, but not scalable for more complex cases.

➡️ {purrr} approach using map(): Link to heading

library(purrr)
numbers <- list(4, 9, 16, 25, -1)
results <- map(numbers, sqrt)

✅ More concise and readable
✅ Handles list structures effortlessly
✅ Easily extendable to more complex functions

📊 Example 2: Handling Errors with possibly() Link to heading

Some operations might fail. {purrr} helps us handle errors gracefully using possibly().

➡️ Handling errors in a safe way: Link to heading

safe_sqrt <- possibly(sqrt, otherwise = NA)
results <- map(numbers, safe_sqrt)

✅ Prevents the entire operation from failing due to a single error

📊 Example 3: Iterating Over Two Lists with map2() Link to heading

We can iterate over two lists simultaneously.

➡️ Multiply corresponding elements in two lists: Link to heading

list1 <- list(2, 4, 6)
list2 <- list(3, 5, 7)

map2(list1, list2, ~ .x * .y)

✅ Efficiently processes paired elements from two lists.

📊 Example 4: Working with Data Frames Using pmap() Link to heading

When we need to apply a function across multiple columns in a dataframe, pmap() is the best choice.

➡️ Example: Compute a BMI column Link to heading

library(dplyr)
df <- tibble(Weight = c(70, 80, 65), Height = c(1.75, 1.82, 1.68))

df <- df %>%
  mutate(BMI = pmap_dbl(list(Weight, Height), ~ .x / (.y^2)))

✅ Computes BMI for each row efficiently.

📊 Example 5: Reducing a List Using reduce() Link to heading

If we need to perform cumulative operations, reduce() is our go-to function.

➡️ Compute the cumulative sum of a list: Link to heading

values <- list(1, 2, 3, 4, 5)
sum_result <- reduce(values, `+`)

✅ Returns the total sum of all values in the list.

📊 Example 6: Using walk() for Side Effects Link to heading

walk() is similar to map() but is used when the function has side effects (e.g., printing, saving files, logging messages).

➡️ Print each element in a list: Link to heading

walk(numbers, print)

✅ Runs functions without returning output, useful for logging or debugging.

📈 Complete Workflow: Functional Programming with {purrr} Link to heading

Let’s put everything together into a cohesive pipeline.

library(purrr)
library(dplyr)

# Sample dataset
df <- tibble(
  Name = c("Alice", "Bob", "Charlie"),
  Score1 = c(85, 90, 78),
  Score2 = c(80, 95, 88)
)

# Apply functions using purrr
df <- df %>%
  mutate(
    Avg_Score = map2_dbl(Score1, Score2, ~ mean(c(.x, .y))),
    Score_Summary = pmap_chr(list(Name, Score1, Score2), 
                             ~ paste(.x, "has an average score of", mean(c(.y, .z))))
  )

✅ Uses map2() and pmap() to efficiently compute new variables

📌 Key Takeaways Link to heading

✅ {purrr} makes iteration in R more intuitive and powerful.
✅ map() and related functions streamline repetitive tasks.
✅ possibly() and safely() help with robust error handling.
✅ map2() and pmap() simplify multi-list operations.
✅ reduce() and walk() offer advanced functional programming workflows.

📌 Next up: String Manipulation Made Easy with stringr! Stay tuned! 🚀

👇 How do you currently handle iteration in R? Let’s discuss!

#Tidyverse #purrr #RStats #DataScience #Bioinformatics #OpenScience #ComputationalBiology