Skip to contents

Using Results for risky computations

using_results.Rmd

Introduction

The Result classes defined in this package can make code more readable and less error-prone as if a function returns a Result object, users of said function are encouraged to consider unhappy paths after all, they might get a Failure. Using Results can abstract away the need for nested if statements and tryCatch blocks, and can facilitate writing composable pipelines of functions that will either end with a success object or a failure object.

Functions that return Result

Writing a function to return a Result signals to users that the function may well fail for a non-exceptional reason:

# A function that might fail
risky_get_stuff_from_server <- function(good = TRUE) {
  if (good) Success(42) else Failure("whoops, couldn't connect!")
}

# Downstream users are now encouraged to inspect the result for success
risky_result <- risky_get_stuff_from_server(good = TRUE)
if (is_success(risky_result)) {
  print(paste0("The answer: ", unwrap(risky_result)))
}
#> [1] "The answer: 42"
#> [1] "The answer: 42"

risky_result <- risky_get_stuff_from_server(good = FALSE)
if (is_failure(risky_result)) {
  cat(paste0("Failed because ", unwrap_fail(risky_result)))
}
#> Failed because whoops, couldn't connect!
#> Failed because cannot find good answer

Wrapping functions

Even if a function returns a Result, it may call other functions that may raise exceptions rather than return Result objects, and so it’s often desired to wrap a function to ensure errors are returned as Failures and successfully computed outputs are returned as Successes.