ZipDiff

Offline side-by-side diff of two roster exports (OneRoster CSV / Savvas Legacy TFF). Shows exactly what a bulk upload would delete.

ZipDiff compares a BEFORE and an AFTER roster export and shows what a differential import would actually do with them — before anyone bypasses a data-threshold guard. Drop two .zip files (or two sets of loose .csv/.txt files) and ZipDiff: • Detects the format automatically: OneRoster v1.2 CSV, OneRoster v1.1 CSV, or the Savvas Legacy TFF format. • Tells bulk files from delta files (from manifest.csv and the rows themselves). Missing rows in a bulk file are deletions; in a delta file only rows marked status=tobedeleted are. • Applies auto-rostering threshold rules: an ingestion is blocked when 50% or more of any file's records would be deleted, or 70% for staff, students and schools during New Year Rollover — which ZipDiff detects from the school-year data and lets you override. • Lists every record that would be deleted, every orphaned reference in the new upload (an enrollment pointing at a user that no longer exists), and the deletion blast radius (everything downstream of a deleted school, user, class or course). • Flags silent delete-and-replace patterns where source IDs changed and the row count barely moved. • Exports a purge-review manifest (Markdown) to attach to a bypass-approval ticket. Privacy: everything runs locally in the extension's own tab. No network access, no permissions, no third-party code, no analytics. Customer data never leaves the machine. Built for teams supporting auto-rostering onboarding; useful for any OneRoster CSV workflow.