Data-Driven Testing

Data-driven testing in xyva separates test logic from test data. Instead of hardcoding values in your specs, you define Variables and Datasets in the Test Hub and bind them to test runs.

Core Concepts

Concept	Purpose	Example
Variable	Single key-value pair injected at runtime	BASE_URL = https://staging.example.com
Dataset	Tabular collection of rows, each row is a test iteration	User table with name, email, role columns
Run Matrix	Cartesian product of selected tests and dataset rows	5 tests x 10 user rows = 50 executions

How It Works

1. Define Variables (env URLs, credentials, flags)
2. Create or import a Dataset (CSV, JSON, or manual entry)
3. Select tests in the Catalog
4. Bind Variables + Dataset to generate a Run Matrix
5. Launch — the Runner iterates through every combination

No Code Changes Needed

Variables are exposed as process.env entries. Your existing process.env.BASE_URL references work without modification.

Variables vs. Datasets

Variables are scalar — one key, one value per run. Use them for environment configuration:

• BASE_URL — target environment
• TEST_USER_EMAIL — login credentials
• FEATURE_FLAG_DARK_MODE — toggle a feature on/off

Datasets are tabular — multiple rows, each creating a separate test iteration. Use them for input variation:

• Login with 20 different user roles
• Form submission with valid and invalid field combinations
• Search queries across multiple languages

Parameterized Execution

When you bind a Dataset to a test selection, xyva generates a Run Matrix:

Dataset "checkout-users" (3 rows) x 2 selected specs
= 6 total test executions

  checkout.spec.ts — row 1 (admin)
  checkout.spec.ts — row 2 (guest)
  checkout.spec.ts — row 3 (blocked-user)
  payment.spec.ts  — row 1 (admin)
  payment.spec.ts  — row 2 (guest)
  payment.spec.ts  — row 3 (blocked-user)

Execution Time

Large matrices multiply execution time. A 50-row dataset bound to 20 specs produces 1,000 runs. Use the Preview Matrix button to review before launching.

Storage

All Variables and Datasets live in .xyva/test-data/ inside your project directory. They are excluded from version control by the default .gitignore entry that xyva creates.

For team sharing, export as JSON and distribute through your preferred channel, or commit the .xyva/test-data/ folder explicitly.

Related Pages

• Variables & Datasets Editor — detailed UI reference
• Test Catalog — selecting tests for data binding
• Runner — executing parameterized runs