How it works
DataPancake® scans 100% of your data, with a standards library such as X12 834 or without one, so the whole structure is found. You shape and configure what it finds; it materializes your data into tables, or Dynamic Tables that refresh continuously as new data arrives, and proves them against the source.
- Data LoaderInventory Source Data with Optional Load/Stage
- Metadata LibraryApply one of our pre-configured standards libraries, such as X12 834, prior to scanning.
- Schema DiscoveryDiscover every attribute path, nested array regardless of depth, stringified JSON, and type variation.
- Schema ShapingIteratively refine the discovered schema, such as converting objects to arrays, or consolidating the schema due to key uniqueness.
- Pipeline DesignConfigure schema metadata including security policy integration, foreign keys, aliases, transformation expressions, virtual attributes and much more.
- Code GenerationGenerate SQL DDL/DML for Dynamic Tables and views with security policy integration. Regenerate any time from the configuration, with nothing hand-edited to maintain or keep in sync.
- DeployDeploy the generated objects through our automated deployment system or use a git integration with your own CI/CD process.
- ReconcileExecute DML reconciliation SQL to produce reconciliation metrics to verify multiple validation lanes including materialization coverage.
- Data Dictionary (LLM Assisted)Generate definitions, synthetic sample values, and synonyms for each dataset, array (child table), and attribute (field).
- Semantic ModelGenerate Cortex-ready YAML or Semantic Views for Natural Language Queries in Snowflake.
Schema Drift MonitoringMonitor new paths, changed types, and structural variation as source data evolves.
Steps 3 and 4 repeat as an iterative cycle, and so do steps 5 to 8. Schema Drift Monitoring leads back to step 3, a continuing cycle.
What could your analytics and AI do with data you can prove is complete?
Ready to try a new approach?
Demo to proof of value to production-ready in a day.