Mastering Debugging in R

1. Core Debugging Concepts

The Debugging Mindset

• Reproduce: Isolate the error conditions
• Locate: Identify where the error occurs
• Diagnose: Understand why it happens
• Fix: Implement and verify the solution

Common Error Types

2. Essential Debugging Tools

Print Debugging

calculate_stats <- function(x) {
  print(paste("Input:", head(x)))  # Debug print
  mean_val <- mean(x)
  print(paste("Mean:", mean_val)) # Debug print
  mean_val / sd(x)
}

traceback()

# After an error occurs:
traceback()
# 3: calculate_stats("text") 
# 2: process_data(input) 
# 1: main_analysis()

browser()

buggy_function <- function(x) {
  browser()  # Pauses execution here
  # Rest of function
}

3. Advanced Debugging Workflows

Interactive Debugging

debug(function_name)  # Sets debug flag
undebug(function_name) # Removes debug flag

# Debug mode allows step-by-step execution:
# n: next line
# c: continue
# Q: quit

Condition Handling

tryCatch(
  {
    risky_operation()
  },
  error = function(e) {
    message("Error occurred: ", e$message)
    NULL
  },
  warning = function(w) {
    message("Warning: ", w$message)
  }
)

4. Environment Inspection

Environment Tools

ls()          # List objects
str(x)        # Examine structure
exists("x")   # Check if object exists
environment() # Current environment

Debugging S3/S4 Methods

debug(print.data.frame)  # Debug S3 method
debugMethods("plot")    # Debug all plot methods

5. Debugging Complex Systems

Parallel Code Debugging

# Debug parallel code by forcing sequential execution
options(mc.cores = 1)

# Use tryCatch in foreach
foreach(i = 1:10) %dopar% {
  tryCatch({
    risky_code(i)
  }, error = function(e) message("Error on ", i))
}

Debugging R Packages

# Load package with debug symbols
devtools::load_all(compile = TRUE)

# Set breakpoints in package code
trace("package_function", browser, at = 3)

6. Specialized Debugging Tools

RStudio Debugging Tools

• Breakpoints (click left gutter)
• Debug pane (step through code)
• Environment inspector

Debugging Packages

library(debugme)    # Lightweight debug messages
library(reactlog)   # For Shiny apps
library(profvis)    # For performance debugging

7. Debugging Prevention

Defensive Programming

validate_input <- function(x) {
  stopifnot(is.numeric(x))
  if (any(x < 0)) warning("Negative values detected")
  invisible(TRUE)
}

Unit Testing

test_that("Input validation works", {
  expect_error(validate_input("text"))
  expect_warning(validate_input(-1:1))
})