rand-engine 0.6.3rc3

Rand Engine v2. Package with some methods to generate random data in different formats. Great to mock data while testing or developing.

🎲 Rand Engine

Generate millions of rows of synthetic data in seconds

High-performance random data generation for testing, development, and prototyping

Quick Start • Features • Examples • Documentation • Benchmarks

---

🎯 What is Rand Engine?

Rand Engine is a Python library that generates realistic synthetic data at scale through simple declarative specifications. Built on NumPy and Pandas for maximum performance.

Perfect for:

• 🧪 Testing ETL/ELT pipelines without production data
• 📊 Load testing and stress testing data systems
• 🎓 Learning data engineering without complex setups
• 🚀 Prototyping applications with realistic datasets
• 🔐 Demos and POCs without exposing sensitive data

---

🚀 Quick Start

Installation

pip install rand-engine

Generate Your First Dataset (3 Lines!)

from rand_engine.main.data_generator import DataGenerator
from rand_engine.examples.common_rand_specs import CommonRandSpecs

# Generate 1 million customer records in seconds
df = DataGenerator(CommonRandSpecs.customers(), seed=42).size(1_000_000).get_df()
print(df.head())

Output:

   customer_id  age           city  total_spent  is_premium registration_date
0    uuid-001    42      São Paulo      1523.50        True        2023-05-12
1    uuid-002    28  Rio de Janeiro       872.33       False        2024-01-08
2    uuid-003    56  Belo Horizonte      4215.89       False        2022-11-23

That's it! You just generated 1 million rows of realistic customer data. 🎉

---

✨ Key Features

🐼 Pandas DataFrames from rand_engine.main.data_generator import DataGenerator df = DataGenerator(spec, seed=42).size(1_000_000).get_df() ✅ All methods (common + advanced) ✅ Correlated columns ✅ Complex patterns ✅ PK/FK constraints

⚡ Spark DataFrames from rand_engine.main.spark_generator import SparkGenerator df = SparkGenerator(spark, F, spec).size(100_000_000).get_df() ✅ Native Spark generation ✅ Databricks ready ✅ Distributed at scale ⚠️ Common methods only

🎁 17+ Pre-Built RandSpecs

No configuration needed! Start generating data immediately:

CommonRandSpecs (Work Everywhere)	AdvancedRandSpecs (Pandas Only)
customers() products() orders()	employees() devices() invoices()
transactions() sensors() users()	shipments() network_devices() vehicles()
	real_estate() healthcare()

# Use any pre-built spec instantly
from rand_engine.examples.common_rand_specs import CommonRandSpecs
from rand_engine.examples.advanced_rand_specs import AdvancedRandSpecs

df_orders = DataGenerator(CommonRandSpecs.orders()).size(50_000).get_df()
df_employees = DataGenerator(AdvancedRandSpecs.employees()).size(1_000).get_df()

📝 Write to Files

# Write to CSV, Parquet, JSON with compression
DataGenerator(spec).size(1_000_000).write() \
    .format("parquet") \
    .compression("snappy") \
    .mode("overwrite") \
    .save("./data/customers")

🌊 Stream Data

# Simulate real-time data streams
DataGenerator(spec).stream() \
    .throughput(min=1000, max=5000) \
    .format("json") \
    .start("./data/stream/events")

---

💡 Usage Examples

1️⃣ Local Development (Pandas)

from rand_engine.main.data_generator import DataGenerator
from rand_engine.examples.common_rand_specs import CommonRandSpecs

# Generate and explore
df = DataGenerator(CommonRandSpecs.transactions(), seed=42).size(100_000).get_df()
print(df.describe())

2️⃣ Databricks / Spark Environments

from rand_engine.main.spark_generator import SparkGenerator
from rand_engine.examples.common_rand_specs import CommonRandSpecs
from pyspark.sql import functions as F

# Generate Spark DataFrame with 100M rows
df_spark = SparkGenerator(spark, F, CommonRandSpecs.orders()).size(100_000_000).get_df()

# Write to Delta Lake
df_spark.write.format("delta").mode("overwrite").save("/path/to/delta/table")

3️⃣ Custom Specifications

# Define your own data structure
custom_spec = {
    "user_id": {
        "method": "unique_ids",
        "kwargs": {"strategy": "uuid4"}
    },
    "age": {
        "method": "integers",
        "kwargs": {"min": 18, "max": 80}
    },
    "salary": {
        "method": "floats",
        "kwargs": {"min": 30000, "max": 150000, "round": 2}
    }
}

df = DataGenerator(custom_spec).size(50_000).get_df()

📖 Learn more: BUILD_RAND_SPECS.md | 50+ Examples

---

📊 Performance Benchmarks

Real-world performance tests across different environments:

Environment	Dataset	Rows	Time	Throughput
Local (Python 3.12)	Customers	1M	81.5s	~12K rows/sec
Databricks (Standard)	Customers	1M	7.4s	~135K rows/sec
Databricks (Spark)	Orders	100M	19.4s	~5.1M rows/sec
Databricks (Custom)	Custom Spec	100M	19.4s	~5.1M rows/sec

💡 Tip: Spark generation scales linearly with cluster size for massive datasets (100M+ rows).

---

🔑 Advanced Features

🔗 Constraints System - Referential Integrity

Generate multiple related tables with Primary Keys (PK) and Foreign Keys (FK):

from rand_engine.main.data_generator import DataGenerator

# Define specs with constraints
customers_spec = {
    "customer_id": {"method": "unique_ids", "kwargs": {"strategy": "sequence"}},
    "name": {"method": "distincts", "kwargs": {"distincts": ["Alice", "Bob", "Charlie"]}},
    "constraints": {
        "pk_customer": {"tipo": "PK", "fields": ["customer_id"]}
    }
}

orders_spec = {
    "order_id": {"method": "unique_ids", "kwargs": {"strategy": "sequence"}},
    "customer_id": {"method": "integers", "kwargs": {"min": 1, "max": 1000}},
    "amount": {"method": "floats", "kwargs": {"min": 10, "max": 1000, "round": 2}},
    "constraints": {
        "fk_customer": {
            "tipo": "FK",
            "fields": ["customer_id"],
            "references": {"spec_name": "customers", "pk_name": "pk_customer"}
        }
    }
}

# Generate with referential integrity
generator = DataGenerator({"customers": customers_spec, "orders": orders_spec})
dfs = generator.size({"customers": 1000, "orders": 5000}).get_dfs()

📖 Complete guide: CONSTRAINTS.md

🎨 Advanced Methods - Correlated Data

Generate correlated columns for realistic patterns:

# Currency-Country correlations  
orders_spec = {
    "order_id": {"method": "unique_ids", "kwargs": {"strategy": "sequence"}},
    "currency_country": {
        "method": "distincts_map",  # Correlated pairs
        "splitable": True,
        "cols": ["currency", "country"],
        "sep": ";",
        "kwargs": {"distincts": ["USD;US", "EUR;DE", "BRL;BR", "JPY;JP"]}
    }
}

df = DataGenerator(orders_spec).size(10_000).get_df()
# Result: USD always paired with US, EUR with DE, etc.

Available Advanced Methods:

• distincts_map - Correlated pairs (currency ↔ country)
• distincts_multi_map - Hierarchical combinations (dept → level → role)
• distincts_map_prop - Weighted correlated pairs
• complex_distincts - Pattern-based strings (IPs, SKUs, URLs)

📖 Complete guide: BUILD_RAND_SPECS.md

---

💡 Quick Tips

🎯 For Data Engineers Use seed for reproducible tests Export to Parquet for large datasets Use constraints for multi-table integrity Stream mode for real-time testing

🧪 For QA Engineers Start with pre-built specs Generate edge cases with probabilities Multiple seeds = multiple test scenarios Test PK/FK relationships

---

📚 Documentation

Document	Description
BUILD_RAND_SPECS.md	Complete guide to building custom specifications
EXAMPLES.md	50+ production-ready examples
CONSTRAINTS.md	PK/FK system and referential integrity
API_REFERENCE.md	Full method reference
LOGGING.md	Logging configuration

---

🧪 Testing

494 tests passing with comprehensive coverage:

pytest                                    # Run all tests
pytest tests/test_2_data_generator.py -v # Test DataGenerator
pytest tests/test_3_spark_generator.py -v # Test SparkGenerator
pytest tests/test_8_consistency.py -v    # Test constraints

---

📦 Requirements

• Python >= 3.10
• numpy >= 2.1.1
• pandas >= 2.2.2
• faker >= 28.4.1 (optional)
• duckdb >= 1.1.0 (optional)

---

🤝 Contributing

Contributions are welcome! Feel free to:

• 🐛 Report bugs via Issues
• 💡 Suggest features via Discussions
• 🔧 Submit pull requests

---

📞 Support

• GitHub Issues: Report bugs
• GitHub Discussions: Ask questions
• Email: [e-mail]

---

📄 License

MIT License - see LICENSE for details.

---

🌟 Star the project if you find it useful!

Built with ❤️ for Data Engineers and the data community

⬆ Back to top