Converting ETL¶
Converting ETL means moving legacy data integration jobs to Snowflake by separating transformation (how data moves between sources and targets) from orchestration (when jobs run, dependencies, and variables). On Snowflake, transformation typically lands in dbt projects on Snowflake and orchestration maps to native Snowflake constructs such as tasks and stored procedures. The Snowflake AIM Agent for Data Warehouses drives this conversion for SSIS and Informatica PowerCenter workloads.
ETL conversion in the Snowflake AIM Agent for Data Warehouses¶
The agent treats ETL as a first-class capability alongside SQL conversion, assessment, and data migration:
| Capability | What it means |
|---|---|
| Convert ETL pipelines | Translate SSIS packages and Informatica workflows toward dbt using deterministic conversion, with optional AI-assisted remediation for remaining gaps |
| Assess workloads | The interactive HTML report includes SSIS / Informatica ETL analysis (package classification, control- and data-flow mapping, effort estimates) alongside waves, exclusions, and dynamic SQL |
| Output layout | Replatform output lands under Output/ETL/ with shared etl_configuration/ plus per-package folders that combine orchestration SQL (tasks or procedures) with per-data-flow dbt projects (staging, intermediate, marts) |
The conceptual mapping from source ETL constructs to Snowflake artifacts looks like this:
| Source concept | Typical output role |
|---|---|
| SSIS Data Flow Task | Standalone dbt project (transformation) |
| SSIS Control Flow | TASK or procedure (orchestration) |
| SSIS variables | control_variables-style infrastructure plus dbt variables |
| Informatica mapping / workflow | dbt models plus orchestration tasks |
Example prompts (from the skill topic):
Platform availability for SSIS to dbt, Informatica to dbt, and the corresponding AI conversion paths appears in the Supported source systems table in the Snowflake AIM Agent for Data Warehouses (deterministic and AI-assisted conversion are available across SQL Server, Redshift, Teradata, Oracle, and other source systems).