← 피드로
Flaky tests are frequently caused by unpredictable or mutated test data. Managing test data deterministically—treating it with the same rigor and version control as application code—ensures reproducible runs across local machines and CI/CD pipelines.
Here is a practical guide on versioning fixtures, seeding environments, and automating test data generation.
Key Recommendations for Deterministic Test Data
- Version Control Your Fixtures: Store canonical, non-sensitive fixtures in your repository (e.g., inside a test-data/ directory) and commit changes via Git.
- Automate Database Seeding: Maintain a data/seeds/ directory and a dedicated script (e.g., scripts/seed-data.js) to reset and seed your target test database directly from versioned fixtures.
- Integrate Generation into CI: Add CI steps to dynamically generate test data files and validate checksums before running test suites.
- Establish Environment Refresh Policies: For shared or mutable test environments, enforce daily snapshot rollbacks or on-demand pipeline jobs to clean and reset state.
Example Workflow
Generate fresh, standardized excel datasets or run bundled data scripts using the following steps:
# Execute local generator script directly
node test-data/excel/generateSampleData.js
# Or invoke via npm scripts
npm run generate:excel-data
Enter fullscreen mode Exit fullscreen mode