The highest-risk export problem is not always a broken file. It may be a file that opens normally but omits a supplier, misstates a quantity, mixes currencies, or repeats a line. Audit the data before using it as a migration source.
1. Recognize the file type from its columns
Do not trust the file name alone. Identify purchase-order, cost, and stocktake exports from their header fields. A purchase-order file should expose PO, supplier, SKU, ordered quantity, received quantity, and cost concepts. A stocktake file should expose location, SKU, expected quantity, counted quantity, and variance concepts.
Flag an unrecognized file instead of forcing it into the wrong schema. Preserve the original file for manual review.
2. Validate required values
For purchase orders, check blank PO numbers, suppliers, SKUs, ordered quantities, and unit costs. For stocktakes, check missing locations, SKUs, expected quantities, and counted quantities. Validate that numeric fields actually contain numbers and that negative values are intentional.
3. Reconcile ordered and received quantities
Calculate ordered minus received quantity for every PO line. Group the results by order and supplier. Review short receipts, over-receipts, and closed or received orders that still show outstanding quantities. These are often the records that become difficult to explain after the original interface disappears.
4. Find duplicate candidates
Group rows using a practical key such as PO number, SKU, and location. Repeated keys are candidates, not automatic errors. Compare dates, quantities, receipt references, and status to distinguish split shipments from duplicates. Keep an issue list showing the row locations and the decision.
5. Review stocktake variances
Recalculate counted quantity minus expected quantity. Compare that result with any exported variance field. Sort by absolute variance to find the largest differences first. Confirm that locations are not accidentally combined and that zero or blank values have not been treated as equivalent.
6. Build cost history by SKU
Sort each SKU's costs by date and flag material changes. A 20 percent threshold is a useful review starting point, not a universal accounting rule. Check whether the change came from a new supplier, currency, pack size, freight method, or a genuine unit-price movement.
7. Protect generated CSV files
Values beginning with characters such as =, +, -, or @ may be interpreted as spreadsheet formulas. Neutralize formula-like text in generated reports so opening the archive does not execute unintended spreadsheet expressions.
8. Record coverage and fingerprints
For every file, record the name, size, data-row count, recognized type, mapped columns, and SHA-256 fingerprint. Add totals for files recognized, POs, suppliers, issues, landed costs, and stocktake lines. The audit should make limitations visible rather than presenting uncertain data as complete.
9. Verify a sample against Stocky
Automated checks identify patterns, but they cannot know every store-specific exception. Compare representative rows with Stocky while access is available. Document discrepancies and retain the raw export alongside cleaned output.
Run these checks automatically
Select up to 10 CSV files. The free scan runs locally in the browser and shows recognition, totals, risk score, and the first five findings.
Audit Stocky CSV files