From packing list
PDF to useful CSV.
A useful CSV needs more than text copied out of a PDF. It needs a consistent structure, explicit units, and a clear link back to the source.
1. Start with the destination columns
Open the import specification or spreadsheet your receiving team already uses. Write down the required column names and the expected type of each value. A basic line-item export might use product_sku, description, quantity, quantity_unit, carton_count, and gross_weight_kg.
Keep shipment-level fields such as purchase_order and packing_list_number on each row if your destination needs them. Confirm the required delimiter and character encoding before preparing a full batch.
2. Keep one source line item per row
Do not flatten a supplier’s entire table into a single description. Preserve separate SKUs, variants, and units. If a packing list groups products by carton, decide whether your destination expects a row per product or a row per carton before combining any values.
Store product codes as text. Spreadsheet software can remove leading zeroes or interpret a code as a date, changing the identifier you need for receiving.
3. Make units explicit
A quantity of 12 could mean 12 pieces, 12 packs, or 12 cartons. Preserve the source unit and confirm any conversion with the supplier or your internal product data. Do the same for pounds versus kilograms and dimensions in inches versus centimetres.
4. Check totals against the source
Add the line-item quantities and compare them with the document total when the units match. Compare carton counts carefully: a mixed carton may contain several product lines, so summing a carton count on every row can double-count the same carton.
Keep an empty weight field empty until you can resolve it. Entering zero can make missing information look like a measured value.
5. Review, then test the import
Check references, product codes, quantities, and exceptions alongside the PDF. Import a small sample into a test environment or a copy of your spreadsheet before using the full file in an operational system.
Retain the source document and record any corrections so a colleague can trace a row back to the shipment paperwork. Treat AI extraction as a draft for review.
Where CartonField fits
CartonField is being developed for the extraction, review, and field-mapping steps. Our static sample shows the intended experience; live processing and CSV downloads are not available yet.