JSON Cleanup with Duplicate and Exception Review

JSON Cleanup with Duplicate and Exception Review

This self-initiated project demonstrates a reviewable JSON cleanup workflow using eight fictional contact and event records. It is a synthetic portfolio sample, not paid client work or a deployed client service.

The tool applies explicit field rules: trimming selected text, normalizing email domains while preserving local-part case, and converting supported timestamps to UTC. String IDs remain exact, including whitespace; a warning highlights IDs that need manual review. Records are matched by kind and exact ID, with the first valid record retained. Identical normalized duplicates are separated, and conflicting or invalid records are quarantined.

The generated result contains 3 cleaned records, 1 identical duplicate and 4 quarantined records: 1 conflicting ID collision and 3 invalid records. Every input row is accounted for. Outputs include a change report, review reasons and an original backup that preserves the input bytes.

Five core Node tests and six independent audit tests passed. Checks cover normalization, exact-ID separation, numeric hazards, malformed input, backup integrity, refusal to overwrite and idempotence: cleaning the cleaned output makes no further changes.

The sample demonstrates JSON processing only. Native JavaScript number parsing has precision limits; the raw backup preserves the original source. It does not establish Excel delivery, large-file streaming, client outcomes or earnings.