Flatfile relies on the mapping decisions of your users to learn what mapping decisions should be expected, and create better predictions for future users that use the importer. We do the following to ensure that mapping can be as accurate as possible and any erroneous mapping decisions can be removed from your mapping history:
We segregate history by environment so that test runs don't impact production.
We only save mapping history after a file has been fully mapped into a destination workbook so history isn't impacted by somebody clicking around in the UI without intending to complete the workflow.
We remember history on a schema-by-schema basis so that all enum history has the proper context of the surrounding file.
We forget past history as soon as a newer option is available.
We offer self-serve endpoints to delete history on demand.