What problem does it solve? Migrating SQL Server Integration Services (SSIS) packages to Databricks requires manually reading opaque .dtsx XML files, inventorying Control Flow tasks and Data Flow components, and mapping each one to a modern equivalent — a slow, error-prone process that often loses error outputs, surrogate keys, or reconciliation checks. ## Core Features & Use Cases - Buffer-safe .dtsx parsing: Extracts executables, Data Flow components, connection managers, variables, and precedence constraints via Python/lxml into a JSON index without dumping raw XML into context. - Complexity classification and mapping: Classifies packages as Simple/Medium/Complex/Blocked and maps each task to Databricks equivalents (Workflows/Jobs, Lakeflow DLT, Delta MERGE/SCD, Auto Loader) using the normative kb/ssis-migration knowledge base. - DAB-packaged generation with reconciliation: Produces Bronze/Silver/Gold layers packaged as Declarative Automation Bundles, runs a mandatory sanity self-review checklist, and reconciles source vs. target row counts and numeric sums. - Use Case: Given a directory of 40 legacy .dtsx packages, generate an inventory table, convert each package into idempotent PySpark/Delta pipelines with quarantine handling, and deliver a conversion report flagging Script Tasks and Fuzzy components for manual review. ## Quick Start Ask the agent to parse all .dtsx files in your SSIS project folder and generate the Databricks migration plan with converted pipelines and a reconciliation report.
