PBI2Snow

Power BI to Snowflake SQL Translator - Convert Power BI models (.bim files) to Snowflake SQL with enhanced M and DAX support.

Features

• Complete Model Translation: Extract and convert entire Power BI data models
• M Language Support: Advanced translation of Power Query M expressions to SQL
• DAX Translation: Convert DAX formulas to Snowflake SQL equivalents
• Relationship Mapping: Preserve table relationships and foreign keys
• View Generation: Create optimized Snowflake views with proper naming conventions
• Configuration Options: Flexible translation modes and confidence levels
• Production Ready: Comprehensive error handling and logging

Installation

From PyPI (Recommended)

pip install pbi2snow

From Source

git clone https://github.com/llmsresearch/pbi2snow.git
cd pbi2snow
pip install -e .

Quick Start

Command Line Usage

# Basic usage
pbi2snow --bim-file model.bim --output-dir output/

# With custom schema and verbose output
pbi2snow --bim-file model.bim --output-dir output/ --target-schema MY_SCHEMA --verbose

# Production mode with high confidence threshold
pbi2snow --bim-file model.bim --mode production --confidence-threshold 80

Python API

from pbi2snow import extract
from pbi2snow.translator import UnifiedTranslator, TranslationConfig

# Load Power BI model
model_data = extract.collect('path/to/model.bim')

# Configure translator
config = TranslationConfig(
    target_schema='SEMANTIC',
    confidence_threshold=70,
    mode='production'
)

# Translate model
translator = UnifiedTranslator(config)
results = []

for table in model_data['tables']:
    result = translator.translate_table(table)
    results.append(result)

print(f"Translated {len(results)} tables")

Command Line Options

Translation Modes

Conservative Mode

• High accuracy, lower coverage
• Only translates well-understood patterns
• Confidence threshold: 80+

Balanced Mode (Default)

• Good balance of accuracy and coverage
• Handles most common scenarios
• Confidence threshold: 50+

Aggressive Mode

• Maximum coverage, experimental translations
• May require manual review
• Confidence threshold: 20+

Output Structure

output/
├── views/                  # Generated Snowflake views
│   ├── V_TABLE_NAME.sql   # Individual view files
│   └── ...
├── manifest.json          # Translation summary and metadata
└── all_views.sql          # Combined SQL file

Supported Power BI Features

M Language (Power Query)

• Table transformations
• Column operations
• Filtering and grouping
• Joins and merges
• Data type conversions
• Custom functions (limited)

DAX Expressions

• Basic calculations
• Aggregations (SUM, COUNT, AVG, etc.)
• Time intelligence functions
• Filter functions (FILTER, CALCULATE)
• Relationship functions
• Complex nested expressions (limited)

Model Elements

• Tables and columns
• Relationships and foreign keys
• Calculated columns
• Calculated tables
• Measures (basic)

Use Cases

• Data Warehouse Migration: Move Power BI models to Snowflake
• SQL Modernization: Convert legacy BI logic to modern SQL
• Multi-Platform Support: Run the same logic on different platforms
• Performance Optimization: Leverage Snowflake's compute power
• Compliance & Governance: Centralize data transformations

Development Setup

# Clone the repository
git clone https://github.com/llmsresearch/pbi2snow.git
cd pbi2snow

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black pbi2snow/
flake8 pbi2snow/

License

This project is licensed under the MIT License - see the LICENSE file for details.

Roadmap

• Enhanced DAX function support
• Azure Synapse SQL support
• Power BI Premium integration
• Real-time translation validation
• Visual Studio Code extension
• Automated testing framework