How to Generate Bulk Test Data for Database Seeding

Why Seed Databases with Test Data?

Development, staging, and test databases need realistic data to function properly. Empty databases do not reveal performance issues, pagination bugs, search indexing problems, or UI layout issues that only appear with real-world data volumes.

Address data is one of the most common types of seeded data because nearly every application that handles users, customers, orders, or locations stores addresses.

How Much Data Do You Need?

Development Environment

Typically 100-1000 records. Enough to:

Staging Environment

Typically 10,000-100,000 records. Enough to:

Performance Testing

100,000+ records. Specifically for:

Generation Approaches

Approach 1: Web-Based Generators

Use a web-based address generator with CSV or JSON export:

Best for: Small to medium datasets (up to 1000 records) when you need data quickly.

Approach 2: Code Libraries

Use a library like Faker to generate data programmatically:

// Node.js example using @faker-js/faker
const { faker } = require('@faker-js/faker');
const fs = require('fs');


const addresses = [];
for (let i = 0; i < 10000; i++) {
  addresses.push({
    id: i + 1,
    street: faker.location.streetAddress(),
    city: faker.location.city(),
    state: faker.location.state({ abbreviated: true }),
    zip: faker.location.zipCode(),
    country: 'US',
    phone: faker.phone.number(),
    created_at: faker.date.past().toISOString(),
  });
}


fs.writeFileSync('addresses.json', JSON.stringify(addresses, null, 2));

Best for: Large datasets and automated seeding pipelines.

Approach 3: SQL Seed Scripts

Generate INSERT statements directly:

INSERT INTO addresses (street, city, state, zip, country) VALUES
('742 Oak Avenue', 'Portland', 'OR', '97201', 'US'),
('1588 Elm Drive', 'Austin', 'TX', '73301', 'US'),
('324 Maple Lane', 'Denver', 'CO', '80201', 'US');
-- ... thousands more rows

Best for: When you want a portable seed file that runs in any SQL client.

Approach 4: ORM Seed Scripts

Most ORMs and frameworks have seeding mechanisms:

// Prisma seed example
async function main() {
  for (let i = 0; i < 1000; i++) {
    await prisma.address.create({
      data: {
        street: faker.location.streetAddress(),
        city: faker.location.city(),
        state: faker.location.state({ abbreviated: true }),
        zip: faker.location.zipCode(),
      },
    });
  }
}

Best for: Projects already using an ORM with built-in seed support.

Multi-Country Seeding

For international applications, distribute your seed data across target countries: