We are building a controlled group-synchronization process in ServiceNow where group records are fetched from another instance (via a REST-based custom data source), loaded into an import set, and then transformed into the sys_user_group table using a transform map with coalescing on sys_id so that existing groups are updated and missing groups are created without duplication. Along the way, we are troubleshooting an issue where extra “ghost” group records (with name = sys_id and other fields empty) are being created, which happens when ServiceNow inserts a target record before field maps are applied—typically due to transform lifecycle behavior such as multiple active transform maps or early insert decisions—so the goal is to ensure only one transform map runs, field mappings apply cleanly, and the sync reliably results in one correct group per source record, no duplicates, and no empty rows.
