Airbyte is an open-source data integration engine that helps you consolidate your data in your data warehouses, lakes and databases.

Airbyte

when using cdc with the postgres connector, every time tehre is an update to a field, the airbyte exctracted at field gets updated and get a new timstamp, correct?

Not exactly. The `_airbyte_extracted_at` (also referred to as `_airbyte_emitted_at` in older versions) reflects *when Airbyte emitted/extracted the record during a sync*, not when the field was updated in the source.

Here's what actually happens with CDC and updates:

• When a record is *updated* in the source, Airbyte captures that change event from the WAL (write-ahead log) and emits a new record to the destination during the next sync. [<https://docs.airbyte.com/platform/understanding-airbyte/cdc|CDC Docs>]
• The new emitted record *will have a new `_airbyte_extracted_at` timestamp*, but only because it is a *newly emitted record* — not because the field itself tracks source-side update times.
• The CDC-specific metadata field `_ab_cdc_updated_at` is what captures the *timestamp of the actual database transaction* that caused the change (INSERT/UPDATE/DELETE). [<https://airbyte.com/tutorials/incremental-change-data-capture-cdc-replication|CDC Tutorial>]
To illustrate: when a record with `id=1` was updated twice in the tutorial example, two new raw records appeared in the destination, each with the same `_airbyte_emitted_at` (since they were captured in the same sync), but with different `_ab_cdc_updated_at` values reflecting the actual transaction times. [<https://airbyte.com/tutorials/incremental-change-data-capture-cdc-replication|CDC Tutorial>]

*In summary:*
• `_ab_cdc_updated_at` → timestamp of the source database change
• `_airbyte_extracted_at` / `_airbyte_emitted_at` → timestamp of when Airbyte synced/emitted the record :thumbsup: 0 button :thumbsdown: 0 button :information_source: *Tip:* Mention *@kapa.ai* in the thread for follow-ups or *DM the app* for *1-to-1* help.

<@U04S8PLJWMT> Okay, but my question is when ther eis a change in a record, cdc would capture it and have an updated record with a newer `_airbyte_extracted_at`

Yes, that is correct. When CDC captures a change to a record, the newly emitted record will have a newer `_airbyte_extracted_at` (or `_airbyte_emitted_at`) timestamp compared to the original record, since it is emitted during a later sync.

You can see this clearly in the tutorial example. After updating the record with `id=1`, the raw table shows the new entries with a later `_airbyte_emitted_at` timestamp:

• *Original record* (id=1, name="A1 CDCSyn"): `_airbyte_emitted_at = 2022-09-26 18:07:07.009+00`
• *Updated record* (id=1, name="D1 CDCSyn"): `_airbyte_emitted_at = 2022-09-26 18:19:31.857+00`
[<https://airbyte.com/tutorials/incremental-change-data-capture-cdc-replication|CDC Tutorial>]

So yes — each time a change is captured and synced, the resulting record in the destination will carry a newer `_airbyte_extracted_at` timestamp reflecting when that sync ran. :thumbsup: 0 button :thumbsdown: 0 button :information_source: *Tip:* Mention *@kapa.ai* in the thread for follow-ups or *DM the app* for *1-to-1* help.