ReleaseBytes
databricks Databricks Blog ·

Databricks Genie Code Beta: Agentic Converter for SQL Dialect Migration

blogdatadatabricksengineer
feature announcement

Databricks has launched the Agentic Converter in Genie Code as a beta feature, utilizing AI agents to automate the conversion of proprietary SQL dialects (T-SQL, Snowflake, Redshift, Oracle, BigQuery, Teradata) to open ANSI SQL. This aims to simplify and accelerate legacy data warehouse migrations to the Databricks Lakehouse by handling code analysis, iterative conversion, and syntax validation. The tool introduces migration projects for tracking progress and lineage, alongside custom skill creation for tailored conversions, making data warehouse migration a configurable and monitorable process.

  • →Agentic Converter in Genie Code for SQL Migration
  • →Migration Projects for Enhanced Management
  • →Custom Skills for Tailored Conversions
  • →Code Complexity Scoring and Lineage Visualization
  • →Support for Enterprise SQL Features
Features (3) ›
  • Agentic Converter in Genie Code for SQL Migration

    The new agentic converter, now in beta, uses AI agents to convert proprietary SQL dialects (T-SQL, Snowflake, Redshift, Oracle, BigQuery, Teradata) to open ANSI SQL. It employs swarms of parallel agents for iterative conversion, validating syntax and semantic intent to preserve original business logic.

  • Migration Projects for Enhanced Management

    A new 'migration project' feature within Databricks workspaces allows teams to manage source files, track conversion progress, visualize lineage, and identify objects that need to migrate together. Projects also provide a central hub for collaboration during the migration process.

  • Custom Skills for Tailored Conversions

    Users can create custom skills within Genie Code to codify conversion rules, enabling broader customization such as conforming to approved ETL patterns or internal naming conventions. This allows for automated application of specific fixes across the codebase.

Enhancements (1) ›
  • Code Complexity Scoring and Lineage Visualization

    Genie Code analyzes and scores the complexity of each SQL file, enabling teams to prioritize simpler files for migration. It also generates lineage graphs showing dependencies between objects, aiding in migration planning and sequencing.

Notes (1) ›
  • Support for Enterprise SQL Features

    Databricks ensures that essential enterprise SQL features like Multi-Statement Transactions, Temp Tables, and Stored Procedures, commonly used in legacy data warehouses, are fully available in Databricks. This eliminates the need to re-engineer logic for compatibility.

Read the original announcement →

https://www.databricks.com/blog/convert-proprietary-code-open-ansi-sql-genie-code

Related releases