tdd-workflow

Test-driven development in R uses testthat to write failing tests before implementing code. You set up testthat in your package, generate a failing unit test, and implement the minimum code required to pass it.

The recommended code coverage target for R unit tests is 80% or higher overall. You must enforce strict coverage expectations specifically for statistical calculations and data validation logic to ensure reliable refactoring.

You organize testthat tests under the tests/testthat directory using fixtures, helpers, and setup files. This package-native test organization supports reproducible checks for unit, integration, and snapshot tests.

Snapshot testing validates R code behavior by capturing expected outputs and comparing them against future runs. You can use snapshot tests alongside unit and integration tests to keep behavior correct and stable as your codebase evolves.

You should apply a test-first workflow when building data processing pipelines, adding new functions, fixing bugs, refactoring packages, extending models, or developing Shiny components to reduce regressions and ensure output validation.

Test-driven development works for fixing bugs in R packages by requiring you to write a failing testthat test that reproduces the bug first. You then implement the fix to make the test pass while tracking coverage toward 80%.