Good afternoon,
among the best practices suggested by Celonis, there are:
- Scheduling Full Loads if delta loads are not possible or if source tables' metadata are changed frequently (i.e., a column is added or deleted).
- Extracting only the relevant columns of the source tables (i..e, those actually used in the transformations).
If the relevant columns can't be deleted or changed in the source systems (because, for instance, they are crucial for the business and it would be a disaster to change them for the data recovery!), why should I schedule full extractions to consider metadata changes?
Shouldn't I launch them only manually when columns need to be added to the extracted tables?
Moreover, if the answer is in the documentation, could you share the link?
Thanks and regards,
Simone