Enable CDC in Lakeflow Connect Tables
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago
Hello everyone, I'm implementing a project that requires AutoCDC workflows using Lakeflow.
The architecture is as follows: Data is ingested from a database using Lakeflow Connect, and then a declarative pipeline performs some transformations on this data to load it into a streaming table.
The declarative pipeline needs to capture the changes made to the streaming table generated by the Lakeflow Connect pipeline. So, how can I enable the `delta.enable.ChangeDataFeed` property on the table generated by Lakeflow Connect? Is there a way to do this? Or do these tables already contain the necessary CDC metadata by default?
Thanks in advance for your answers.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago - last edited 3 weeks ago
Hi @Nmtc9to5 ,
You likely don't need to manually enable delta.enableChangeDataFeed on Lakeflow Connect tables for most use cases, as they typically contain the CDC metadata by default. (docs)
You can check if it is enabled on the table with SHOW TBLPROPERTIES catalog.schema.your_table;
Use STREAM() in your downstream declarative pipeline to incrementally read from the Lakeflow Connect table.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago
The short answer is no, you don't need to enable it, because the tables generated by Lakeflow Connect already capture and stream CDC metadata natively. In fact, streaming tables in Delta Live Tables (DLT) and Lakeflow pipelines are built on top of an append-only transaction log mechanism that naturally tracks incremental row changes (inserts, updates, deletes) without needing standard CDF turned on.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago
Lakeflow Connect tables generally already include the CDC metadata required for downstream processing. I'd first check:
SHOW TBLPROPERTIES <catalog>.<schema>.<table>;