Skip to content

Spindle in 60 Seconds

Spindle generates realistic, multi-table synthetic data for Microsoft Fabric and beyond.

Install

pip install sqllocks-spindle

Generate

from sqllocks_spindle import Spindle, RetailDomain

result = Spindle().generate(domain=RetailDomain(), scale="small", seed=42)
print(result)
# 9 tables: customer (1000), address (1500), product (500), order (5000), ...

Use

# DataFrames — ready for Pandas, Spark, or any pipeline
customers = result["customer"]
orders = result["order"]

# Write to files
result.to_parquet("./output/retail")
result.to_csv("./output/retail")

# Write SQL (Fabric Warehouse compatible)
sql = result.to_sql_inserts(dialect="tsql", include_ddl=True)

CLI

spindle generate retail --scale small --format parquet --output ./output/
spindle generate healthcare --scale medium --format sql --sql-dialect tsql
spindle list                    # See all 13 domains
spindle describe retail         # Inspect tables, columns, relationships
spindle stream retail --table order --max-events 1000

13 Domains

Retail, Healthcare, Financial, HR, Education, Insurance, IoT, Manufacturing, Marketing, Real Estate, Supply Chain, Telecom, Capital Markets

Key Features

  • Schema-aware — Foreign keys, business rules, realistic distributions
  • 13 calibrated domains — 90+ tables with real-world ratios
  • Multiple scales — demo to xlarge (5K to 100M+ rows)
  • Fabric-native — Lakehouse, Warehouse, SQL Database, Eventstream, Semantic Model
  • Star schema + CDM — Transform 3NF to analytics-ready or Common Data Model
  • Streaming — Real-time event generation with Poisson timing
  • Chaos + Simulation — File drops, corruption, validation gates
  • Reproducible — Same seed = same data, every time

Next Steps