Troubleshooting common parse errors

Five issues account for most parse failures. Each one is fixable in your mapping or the source file.

Missing required column

The file is missing one of the five required canonical fields (po_number, line_no, item_code, quantity, unit_price). Fix by renaming the source column in your mapping, or by adding a Concat manipulator to build the canonical field from two source columns.

Ambiguous date format

A date column contains values like 03/04/2026 that could be either March 4th or April 3rd. ProcuLink will not guess. Attach a DateFormat manipulator that pins the input format explicitly, e.g. dd/MM/yyyy.

UTF-8 BOM and encoding

CSV files exported from Excel sometimes include a UTF-8 byte-order mark, which makes the first header look like po_number. Re-export from Excel as CSV UTF-8 (Comma delimited) (.csv), or open the file in a text editor and remove the leading bytes.

Scanned PDFs

ProcuLink reads text-based PDFs directly — the high-confidence primary path. Image-only scans (phone photos, photocopier output) have no text layer, so they fall back to AI vision extraction; because there is nothing to verify the numbers against, every line is flagged for review before delivery. For the most reliable result, ask the buyer for a text PDF. If a scan is too low-quality to read at all, it will fail — re-scan at a higher resolution or request a text PDF.

European decimal commas

Files that write 12,50 for twelve-and-a-half are handled automatically — semicolon-delimited CSVs are treated as European formatting, and Excel numbers are read as raw values rather than display text. If quantities or prices ever look off by 10× or 100× after a parse, don't send — flag it to support with the file attached.

Supplier item code mismatch

A line parses but its item_code does not match any SKU in the supplier mapping. Either add the SKU to the supplier mapping, or accept an AI suggestion on the Review screen when one is offered.

Where failures queue up

Every parse failure (and every later failure — transform, delivery, supplier rejection) opens an entry in Operations → Exceptions, so nothing fails silently. See working the exceptions queue for how to triage that list and what to do when an order looks stuck rather than failed.

Need help? Email support@proculink.eu or see PO field mapping basics.