Product duplicate audit with unchanged source records

Self-initiated synthetic-data demonstration / AI-assisted. No client data.

This report-only study audits 20 original product records without changing, deleting or merging source data. It distinguishes exact duplicate keys and fields from conflicting duplicate keys. All findings link back to their original source-row identifiers.

The defined checks identify three duplicate-key groups involving seven rows: two groups have matching source fields, and one group has a price conflict. Three missing-required-field events and two ambiguous-date events are also reported. There are 12 issue events affecting 12 rows; eight rows have no issue under these specific checks.

Deliverables are the unchanged source CSV, a row-level issues CSV, documented rules and a verification manifest. Original zero-stock values and ambiguous dates remain unchanged for source-owner decisions.

Artifact Tool CSV text roundtrip and independent Python checks covered source preservation, group membership, issue-source links, counts and file hashes. This is a limited demonstration, not a certification of business-record accuracy or a past client engagement. No Microsoft Excel desktop test was performed.