CDM Export¶

Export Spindle-generated data to Microsoft Common Data Model (CDM) format, compatible with Dataverse, Power Apps, and Azure Synapse Analytics.

Quick Start¶

from sqllocks_spindle import Spindle, RetailDomain
from sqllocks_spindle.transform.cdm_mapper import CdmMapper

domain = RetailDomain()
result = Spindle().generate(domain=domain, scale="small", seed=42)

mapper = CdmMapper(
    cdm_map=domain.cdm_map(),
    schema=domain.get_schema(),
)
mapper.export(result.tables, output_dir="./cdm_output/retail")

This produces:

cdm_output/retail/
├── model.json
├── Contact/
│   └── Contact.csv
├── Product/
│   └── Product.csv
├── SalesOrder/
│   └── SalesOrder.csv
└── ...

How It Works¶

Each Spindle domain provides a cdm_map() method that maps source tables to CDM standard entity names:

domain = RetailDomain()
cdm_map = domain.cdm_map()

# Inspect the mapping
for source_table, cdm_entity in cdm_map.mappings.items():
    print(f"{source_table} → {cdm_entity}")

Output:

customer → Contact
address → CustomerAddress
product_category → ProductCategory
product → Product
store → Store
promotion → Campaign
order → SalesOrder
order_line → SalesOrderProduct
return → ReturnOrder

Available CDM Mappings¶

All 13 Spindle domains include CDM mappings:

Domain	CDM Entities
Retail	Contact, CustomerAddress, Product, Store, Campaign, SalesOrder
Healthcare	Patient, Practitioner, Account, Appointment, Condition, Invoice
Financial	Contact, Account, FinancialAccount, Transaction, Loan, Payment
HR	BusinessUnit, Position, Worker, Compensation, LeaveRequest
Education	BusinessUnit, Course, Contact, CourseEnrollment, Award
Insurance	Worker, Contact, Contract, Case, Payment, Assessment
IoT	Category, Location, Asset, Component, Observation, Alert
Manufacturing	BusinessUnit, Product, WorkOrder, QualityOrder, Asset
Marketing	Campaign, Contact, Lead, Opportunity, Activity, Goal
Real Estate	Location, Worker, Listing, Transaction, Appointment
Supply Chain	Warehouse, Vendor, Product, PurchaseOrder, Shipment
Telecom	Product, Contact, Subscription, Invoice, Payment
Capital Markets	Account, Organization, Category, Observation, Transaction

Output Formats¶

The CDM exporter supports multiple file formats within the CDM folder:

mapper.export(
    result.tables,
    output_dir="./cdm_output/retail",
    format="parquet",  # csv (default) | parquet | jsonl
)

model.json¶

The generated model.json manifest follows the CDM folder structure specification:

{
  "name": "retail",
  "version": "1.0",
  "entities": [
    {
      "name": "Contact",
      "source": "customer",
      "attributes": [...]
    }
  ]
}

CLI¶

spindle generate retail --scale small --format cdm --output ./cdm_output/retail

Use Cases¶

Dataverse import — Load CDM entities into Power Platform environments.
Synapse Analytics — CDM folders are natively supported as linked datasets.
Cross-system testing — Validate data pipelines that consume CDM format.
Power Apps prototyping — Populate test data matching Dataverse entity schemas.