When using a data platform like Databricks; is it acceptable to source datasets from the EDW/LakeHouse for operational pipelines? e.g. for a data sync to a third-party platform.