# Invoice 2026/09/014

**Seller:** EPSI sp. z o.o., ul. Bura 21, 00-000 Burowce, NIP PL9998877666
**Buyer:** Northwind Logistics GmbH, Hafenstraße 8, 22222 Hamburg, USt-IdNr. DE881188118

**Issued:** 2026-09-03    **Delivered:** 2026-08-31    **Payment due:** 2026-09-17

## Services rendered

```vmark #lines from example-invoice-csv-import.csv labelled Item, Unit, Qty, Rate [PLN], Net [PLN], VAT [PLN], Gross [PLN] at sha256:e6667fb8dda6801721d885c259e2456a9e6c92f1ec89af7f14330e0c54c6ac14
net_total   [PLN] = SUM(Net)
vat_total   [PLN] = SUM(VAT)
gross_total [PLN] = SUM(Gross)
```

net_total   [PLN] = SUM(Net)
vat_total   [PLN] = SUM(VAT)
gross_total [PLN] = SUM(Gross)

Net of tax the engagement comes to **23300.00**<!--vmark=lines.net_total--> PLN.
VAT at 23% adds **5359.00**<!--vmark=lines.vat_total--> PLN, giving a total due of
**28659.00**<!--vmark=lines.gross_total--> PLN.

---

## Appendix — how this differs from the inline version

This is [example-invoice.md](example-invoice.md) with the `#lines` sheet's rows
moved out to [example-invoice-csv-import.csv](example-invoice-csv-import.csv)
and declared with a `from` clause instead of a preceding GFM table. Everything
else about the sheet is unchanged: `net_total`, `vat_total` and `gross_total`
are the same `SUM` reduces over the same column names, and the anchored
numbers below the block are the same values.

Three things are different because the sheet is now imported rather than
inline:

**No column rules.** The original document computes `Net`, `VAT` and `Gross`
per row inside the `vmark` block (`Net = Qty * Rate`, and so on). An imported
sheet has no cell for a column rule to write to — its columns are read-only
inputs taken straight from the CSV header row — so `Net`, `VAT` and `Gross`
are already computed values *in the CSV*, produced by whatever generated it,
not by VisiMark. Only the aggregate reduces (`SUM`) still run here.

**The `at sha256:…` stamp.** The clause on the fence line pins the CSV's exact
bytes. `visimark check` recomputes the digest on every run and fails with
`STALE` if the file has changed underneath the document, or `IMPORT` if the
clause is missing entirely — an inline table has no analogous guard, because
its data lives in the same text `check` is already reading. Editing
`example-invoice-csv-import.csv` and running `visimark fmt` rewrites only this
`at` clause; the CSV itself is never touched by any VisiMark command.

**`labelled` asserts the header.** `labelled Item, Unit, Qty, Rate [PLN], Net
[PLN], VAT [PLN], Gross [PLN]` checks the CSV's header row matches this exact
name and order before anything is evaluated, so a column renamed or reordered
upstream is a loud `IMPORT` finding instead of a silent misread.

**Units come from the declaration.** The CSV header row stays `Item,Unit,Qty,
Rate,Net,VAT,Gross`, so its bytes and the `sha256` stamp are untouched; the
brackets in the `labelled` list attach `PLN` to the money columns. A CSV that
carried `Rate [PLN]` in its own header would work too, and a header and a
declaration that disagree is a `UNIT` finding. The three totals then declare
`[PLN]`, which `check` verifies against the unit the `SUM`s derive.

See `docs/design/declared-local-data-imports-spec.md` for the full syntax and
semantics.

## Attachment: example-invoice-csv-import.csv

```csv
Item,Unit,Qty,Rate,Net,VAT,Gross
Discovery workshop,day,2,1800.00,3600.00,828.00,4428.00
Backend implementation,hour,64,220.00,14080.00,3238.40,17318.40
Code review retainer,month,1,2500.00,2500.00,575.00,3075.00
On-call support,hour,12,260.00,3120.00,717.60,3837.60
```

Item,Unit,Qty,Rate,Net,VAT,Gross
Discovery workshop,day,2,1800.00,3600.00,828.00,4428.00
Backend implementation,hour,64,220.00,14080.00,3238.40,17318.40
Code review retainer,month,1,2500.00,2500.00,575.00,3075.00
On-call support,hour,12,260.00,3120.00,717.60,3837.60