We would like to request support for dynamic, rolling date range filters on connector tables (for example, current_date - 30 or current_date - 7) that apply incrementally rather than as a static filter.

Currently, when we apply or modify a date filter on a table (as shown in the attached example),

Fivetran triggers a table-level historical re-sync. In a recent instance, this resulted in a re-sync that ran for approximately 6 hours, extracting around 460 GB of data.

Our objective is to configure a dynamic date range filter that:

• Automatically adjusts over time (rolling window)

• Syncs only data within the defined range

• Does not trigger a historical re-sync when the filter definition changes

• Applies incrementally rather than statically
  
  

Business Impact / Justification

• Cost Optimization: Avoiding unnecessary historical re-syncs would significantly reduce data extraction and compute costs associated with large-volume tables.

• Improved Data Freshness: Eliminating long-running re-syncs ensures that incremental data continues to be delivered to end users without delays.

• Operational Efficiency: Dynamic rolling filters would reduce manual intervention, scheduling coordination, and the need for off-hours maintenance windows.

• Scalability: As data volumes grow, static filters that trigger full table re-syncs become increasingly impractical and disruptive.