Preserve the unit associated with an invoice quantity

Where the extra work starts

Keep the source unit with the amount so the next spreadsheet user does not have to reopen every invoice to interpret it.

Upload a representative supplier invoice, inspect its extracted fields and line items beside the PDF, then download Excel. Put the awkward invoice in your trial too: the useful result is less total work after corrections.

This is a transcription and context check. Any conversion or commercial interpretation should use an explicit supported rule and the responsible review, with raw values retained separately.

A practical way to get it right

Inspect the quantity heading, unit column, description and pricing note. Record where the unit is stated and whether it applies to the whole table or a particular line. Keep the exact source wording before considering a standardized code.

For each row, link raw quantity, raw unit and price basis to the same source occurrence. If the unit is absent or ambiguous, mark it unresolved rather than defaulting to “each.” If a description says “box of 12,” do not assume that the quantity column counts boxes without checking the source context.

When a downstream calculation needs normalized units, create separate derived fields with the conversion factor, source evidence and reviewer decision. Never overwrite the source quantity and unit, since the conversion may be wrong or inappropriate for another use.

OpenPeppol: Invoiced quantity associates quantity with a unit-of-measure code in the structured model. Preserve the corresponding unit relationship when transcribing a PDF.

What this looks like on a real task

Synthetic example. The people, references and figures below are invented to demonstrate the method.

The first two rows have equal numbers but different units. P13 cannot safely become 72 items from the description alone. Copy the quantity, unit and state fields into a fixture and include the source location. The examples are invented and do not define a standard packing convention for any supplier.

Catch the cases that create rework

The price may apply to more than one unit, so quantity times printed price can be an invalid check. Preserve any base quantity or “per” wording and ask the reviewer to establish the appropriate comparison. A missing unit may also be stated in a table note; inspect that note before declaring it absent.

If extraction separates the unit from the value or drops it from a description, record the correction and inspect similar rows.

OpenPeppol: Invoice line price details distinguishes price base quantity from price amount. Inspect that basis before applying a simple quantity-times-price calculation.

Decide whether the change is worth keeping

Accept the register only when each relevant quantity has source-backed unit context or an explicit unresolved state. Test the export route to ensure unit strings survive alongside values. If normalized fields are supplied, check them independently and label them derived. The source-preserving register remains useful even when a conversion cannot yet be justified.