Row-Level Ownership: Why One Sync Job Treats 154 and 15 Records Differently

A developer running a nightly data sync for an Illinois workforce development system applies different write rules depending on whether a human has ever edited a record. For 154 untouched directory entries pulled from a state-published feed, the sync overwrites data automatically, since the feed is the only source of knowledge about those rows. For 15 partnered centers where staff have added notes and visit history, any incoming change is queued for human review rather than applied directly. The system also never deletes records, instead stamping a 'missingSince' date to guard against truncated or erroneous exports. The developer argues that write authority should belong to the individual row, not the feed, and has since applied the same two-door principle across an email triage agent, a publishing pipeline, and a medical billing denial engine.