Databricks Lakeflow Introduces REPLACE USING Flows for Pipelines (Beta)
Databricks has launched REPLACE USING flows in Beta for Lakeflow pipelines, simplifying data ingestion from partial snapshots. These flows replace rows matching specified key columns upon update, ordering them by a SEQUENCE BY column to handle out-of-order arrivals. This eliminates the need for change data capture feeds with explicit operations or primary keys, streamlining complex data integration scenarios.
Features (1) ›
- REPLACE USING flows for partial snapshot processing
REPLACE USING flows are now available in Beta for Lakeflow pipelines. They allow users to process partial snapshots, replacing only rows that match specified key columns while leaving others unchanged. Updates are ordered using a SEQUENCE BY column, ensuring correct application even when data arrives out of sequence without requiring CDC feeds or primary keys.
https://docs.databricks.com/aws/en/release-notes/product/2026/august#replace-using-flows-for-pipelines-beta
Related releases
- Zerobus Ingest enabled by default for Databricks compliance security profiles Databricks Release Notes ·
- JAR Tasks on Serverless Compute Now Generally Available on Databricks Databricks Release Notes ·
- JAR Tasks on Serverless Compute Now Generally Available in Databricks Lakeflow Databricks Release Notes ·
- Change Data Feed Now Available for Databricks Materialized Views (Beta) Databricks Release Notes ·
- Databricks enhances Lakeflow pipelines with continuous jobs Databricks Release Notes ·
- Databricks Lakeflow Connect adds SendGrid connector (Beta) Databricks Release Notes ·