Connector

Follow the source instead of reloading it

Full reloads get slower as the source grows and go stale between runs. Change data capture keeps the dataset in step by applying only what changed.

How it works

Two modes

Polling mode compares against a change column. Log mode reads the database replication slot for a complete, ordered change stream.

Inserts, updates and deletes

A delete at the source becomes a delete in the dataset, rather than a row that lingers because the reload query never saw it.

Ordered application

Changes are applied in the order they happened, so an update followed by a delete does not resurrect a row.

Lineage preserved

The dataset keeps its history through the sync — CDC does not reset what you know about it.

What happens after it lands

Once rows are in the catalogue, the connector that produced them stops mattering. The dataset is profiled, can be cleaned, modelled, put in a flow, governed and published exactly like every other dataset — which means switching how data arrives never means rebuilding what happens to it afterwards.

Questions about this connector

What is the difference between polling and log mode?

Polling compares a timestamp or sequence column on a schedule and needs no special database privileges. Log mode reads the replication slot, catches deletes and out-of-band changes that polling misses, and is a Pro capability.

Do I need database admin rights?

Polling mode needs only read access. Log mode needs replication privileges on the source database.

What happens if the sync falls behind?

The lag is reported. In log mode the changes are queued at the source, so catching up applies them in order rather than losing them.

See it on your own data

DataLens is in private beta. Bring a file, a database or an API and work through the whole lifecycle in one sitting.

Request beta access