Videos
In part two of the Databricks Materialized View demo series, Green Leaf Data Architect, Jessica, takes a closer look at refresh behavior after real data changes.
In this video, we test multiple scenarios including single-row updates, multi-million-row updates, small date-based changes, and append-only inserts to see how Databricks handles materialized view maintenance. Along the way, we review performance metrics, scan pruning, event log details, and optimizer-driven decisions between row-based maintenance and complete recompute.
If you want to understand how Databricks materialized views behave in real-world refresh scenarios, this demo walks through the details step by step.
Topics covered:
- Incremental refresh after source table updates
- Row-based maintenance behavior
- Append-only refresh behavior
- Performance observations from bytes read and bytes written
- Event log and planning details
- Optimizer decision-making for refresh strategy
Part one covered the foundational concepts. This follow-up focuses on what happens when those concepts are put into action.