# Data Change Data Feed and Incremental Table Reads
Status: public
Confidence: medium (0.725) (verified)
Last verified: 2026-06-02
Generation: ai_structured


## TL;DR

Change data feeds and incremental table reads help agents process only changed records instead of rescanning an entire lakehouse table.

## Core Explanation

Data pipelines often need to update indexes, features, warehouses, and caches from table changes. Change data feeds and incremental reads expose inserts, updates, and deletes relative to table versions or commit times.

Agents should inspect enablement flags, start version or timestamp, commit retention, delete representation, schema changes, checkpoint state, and whether downstream consumers are idempotent.

## Source-Mapped Facts

- Delta Lake documentation describes Change Data Feed as recording row-level changes between versions of a Delta table. ([source](https://docs.delta.io/delta-change-data-feed/))
- Databricks documentation describes Delta change data feed as enabling tracking row-level changes between versions of a Delta table. ([source](https://docs.databricks.com/aws/en/delta/delta-change-data-feed))
- Apache Hudi documentation describes incremental queries that fetch changed records from a given commit time. ([source](https://hudi.apache.org/docs/sql_queries/))

## Further Reading

- [Delta Lake Change Data Feed](https://docs.delta.io/delta-change-data-feed/)
- [Databricks Delta Change Data Feed](https://docs.databricks.com/aws/en/delta/delta-change-data-feed)
- [Apache Hudi SQL Queries](https://hudi.apache.org/docs/sql_queries/)