Fictional catalog cleanup: 14 rows reconciled

This is an original fictional demonstration, created with AI-assisted production and review. It is not commissioned client work, real inventory or evidence of sales. All source records are invented and contain no personal data.

The task Two supplier extracts use different column names. Fourteen rows include exact repeats, a conflicting SKU, a missing SKU, an invalid price, a blank quantity and incompatible currency/unit values. The goal is one reviewable catalog without inventing missing information.

Rules before cleanup For this exercise, both extracts use the same retailer-assigned SKU namespace. SKUs stay as text, including leading zeros. Only valid USD/EA rows are eligible. Zero quantity is valid. Exact mapped repeats point to the retained source row; their quantities are not added together. Both rows of a conflicting SKU stay on hold. Real work needs buyer-approved mapping and duplicate rules.

What changed 14 source rows = 5 retained + 2 exact repeats + 7 held. The held records include both conflicting 000103 rows, the missing identifier, invalid price, blank quantity, EUR row and PACK row. The exceptions sheet states the decision needed for each item. Every original row remains visible with its source file and CSV line.

Reconciliation Ten rows have comparable valid numeric values in USD/EA. Their quantity reconciles as 68 = 27 retained + 18 repeated + 23 held. Extended amount reconciles as $631 = $294 + $180 + $157. Four rows are explicitly outside numeric coverage. These are illustrative arithmetic totals, not revenue, profit or accounting valuations.

Checks and limits All four workbook audit differences are zero. Export checks verified 84 original source cells, text SKUs, 58 formulas and the exceptions. A separate implementation derived the same results from the source files and rules; expected totals were visible in the rules, so that review was independent but not blind. Native Excel opening and any storefront import were not tested. Classification is a snapshot; changed inputs require a fresh pass.

Handoff The sample includes cleaned CSV/XLSX, the untouched fictional inputs, row lineage, exceptions, mapping rules and an operational summary. Screenshots above show the catalog, exceptions, reconciliation and original rows. A review copy is available on request for a relevant scope discussion.

For a similar small non-personal catalog or inventory task, see the $250 proposed pilot service. Send a short non-sensitive description first; suitability, scope, timing and processing permissions are agreed before starting.

Pilot service: https://contra.com/s/X7eX7OLM-catalog-csv-cleanup-with-exceptions-and-reconciliation