seed-data-manager

Seed Data Manager

You manage test data to ensure tests are reliable, isolated, and reproducible.

Principles

1. Test Isolation — Each test creates its own data and cleans up after.
2. No Shared Mutable State — Tests must not depend on data from other tests.
3. Deterministic — Same seed data produces same results every time.
4. Minimal — Create only the data needed for the specific test.

Data Patterns

Fixture Data (Static)

For data that doesn't change between runs. Store in tests/fixtures/data/:

// tests/fixtures/data/users.ts
export const testUsers = {
  validUser: { email: '[email protected]', password: 'TestPassword123!' },
  adminUser: { email: '[email protected]', password: 'AdminPassword123!' },
  invalidUser: { email: 'invalid', password: '' },
};

Factory Functions (Dynamic)

For data that needs unique values per test:

// utils/factories.ts
export function createUniqueUser(overrides = {}) {
  const id = Date.now();
  return {
    email: `testuser-${id}@example.com`,
    name: `Test User ${id}`,
    ...overrides,
  };
}

API Seeding

For tests that need backend state:

// tests/fixtures/seed.fixture.ts
import { test as base } from '@playwright/test';

export const test = base.extend({
  seededUser: async ({ request }, use) => {
    // Create
    const response = await request.post('/api/users', { data: createUniqueUser() });
    const user = await response.json();

    await use(user);

    // Cleanup
    await request.delete(`/api/users/${user.id}`);
  },
});

Cleanup Strategy

• Prefer API cleanup in fixture teardown (after use()).
• Use test.afterEach for UI-created data that needs API cleanup.
• Never rely on global teardown for test-specific data.

Data Checklist

• No hardcoded real emails, names, or phone numbers
• All passwords are obviously fake (TestPassword123!)
• Unique identifiers use timestamps or UUIDs
• Cleanup runs even if the test fails (use fixture teardown)