Incremental

If you enable synchronization when creating the pipeline, for targets that support incremental load, AutoSync does a full load followed by incremental loads (upserts). The Incremental load type requires that each table have a primary key and a timestamp that is updated each time a change is made to a record. For some sources, relational databases for instance, AutoSync requires you to select the column that contains the timestamp. If you do not select a timestamp field for these sources, AutoSync performs full loads when synchronizing.

If your endpoint doesn't already have a column or field to store the last updated timestamp, you will need to add one. Make sure that the column isn't nullable, contains no null values, and updates the timestamp for all changes. If you can't add a timestamp, you can use Full load instead of Incremental.

For the last updated timestamp field, AutoSync supports values up to nanoseconds. The MySQL destination endpoint is an exception because it doesn't support fractional seconds. The supported data type for the timestamp field differs by endpoint. Refer to the endpoint specific documentation for details.

For most sources, deletions aren't synchronized. However, for Salesforce, you can configure AutoSync to track deletions.

The Incremental load type is supported for: Marketo, Microsoft Dynamics 365 for Sales, Oracle, MySQL, Salesforce, ServiceNow, and SQL Server sources when synchronizing to Azure Synapse, Databricks, Google BigQuery, Oracle, SQL Server, and Snowflake targets.