Language:JapaneseEnglish

oracle-ai-data-platform-workbench-databricks-migrator Claude Code Plugin

Author
Oracle
Category
Development
Topics
Migration & Modernization · Data Engineering & Analytics · Cloud, Deployment & CI/CD
License
MIT(opens in a new tab)
Version
0.2.0
First cataloged
2026-07-09 (UTC)
Explanation last updated
2026-10-16 (UTC)
Source (GitHub) last updated
2026-10-14 (UTC) (21 days ago)

The explanation below is AI-generated. Please verify it against the sources.

This plugin is a Claude Code plugin that lets a user drive migration of Databricks workloads to Oracle AI Data Platform (AIDP) using natural-language requests. According to the README, it bundles a self-contained Python migration engine that first analyzes dependencies and rewrites code (Pass-1), then executes each notebook cell on a live AIDP cluster, verifies the result, and attempts automatic fixes (Pass-2). It also migrates catalog metadata (Unity Catalog / Hive Metastore) into AIDP-compatible DDL using a defined set of rewrite rules and remaps storage paths from Databricks to OCI object storage. The README lists prerequisites for actually running a migration, including installing Python dependencies, OCI authentication, an active AIDP cluster, and an Anthropic API key. It provides a set of skills, slash commands, and reviewer agents covering different phases of the migration (planning, execution, catalog work, and verification).

About the service

Oracle AI Data Platform is described on its homepage as a service for cataloging, ingesting, and analyzing organizational data, offering a platform and framework for building data analytics pipelines. This plugin targets AIDP as the migration destination, connecting to a live AIDP cluster to execute, verify, and fix migrated notebook code and to replay catalog DDL.

What you can do with oracle-ai-data-platform-workbench-databricks-migrator

  • Build a migration manifest/DAG from a Databricks workspace path
  • Scan source tables and paths for availability before starting a migration
  • Run a full job migration combining dependency rewriting and cell-by-cell execution with verification on an AIDP cluster
  • Rewind and re-run a specific failing cell with a stated reason
  • Resume an interrupted migration run
  • Migrate Unity Catalog / Hive Metastore DDL into AIDP using rule-based rewrites
  • Configure bucket/path mappings between Databricks storage and OCI object storage
  • Set up acceptance checks for batch or streaming convergence
  • Use guided slash commands for job migration, catalog migration, data checks, and status summaries
  • Use specialist agents to analyze source notebooks and review migrated output

Sources

Back to list