Skip to content

Setting up Lakeflow Connect to ingest from enterprise sources reliably

Lakeflow Connect provides managed, native connectors for ingesting data from enterprise sources (SaaS applications, databases via CDC, file systems) directly into Unity Catalog-governed Delta tables, eliminating the need for custom ingestion pipelines. It handles schema evolution, incremental sync, and governance automatically, so learners must know which source types are supported and how ingestion pipelines integrate with Unity Catalog's permission model.

Must-know

  • Lakeflow Connect ingestion pipelines write directly into Unity Catalog-managed Delta tables, so target catalogs, schemas, and tables must exist within a Unity Catalog metastore before configuring ingestion.
  • Connectors support incremental ingestion patterns, including change data capture (CDC) from databases, so downstream tables reflect ongoing source changes without full reloads.
  • Ingestion pipelines run as governed workloads, meaning access to both source connections and destination tables is controlled through Unity Catalog permissions and, for source credentials, through securable connection objects.
  • Schema evolution (new columns or type changes at the source) is handled automatically by the connector, but structural changes should still be validated against downstream consumers before promotion.
  • Configuring a connection requires source-specific authentication (e.g., OAuth, API tokens, or database credentials) stored securely and referenced by the pipeline rather than hardcoded.
  • Ingested tables inherit lineage and audit tracking within Unity Catalog, enabling data stewards to trace table origin back to the source connector and pipeline run.
Check this objectiveFree · always available

A data engineer is setting up a Lakeflow Connect ingestion pipeline to bring data from a Salesforce source into a Unity Catalog schema. The source table has a reliable 'LastModifiedDate' field, and the engineer wants new and changed records captured automatically without manually re-ingesting the full dataset each run. Which ingestion approach should they configure?

Your objective map0 tried · 0 right · 33 untouched

What you have tried across Databricks DEA's objectives, not a readiness score.

Coverage checked against the published exam guide on Jul 25, 2026.

These are independent practice questions, written against this certification's published exam guide. They are not the certification vendor's own questions, and not the real exam.