ShrinkWatch EU 如何辨識 20 多種歐盟語言的單價標籤

EU law (directive 98/6/EC) says shops must show a unit price: the price per kilo, per litre or per piece. Every country calls it something different, and every shop prints it a little differently. My open-source browser extension ShrinkWatch EU computes the unit price itself from the pack size and price, then cross-checks it against the shop's own figure. To do that cross-check it has to find the shop's figure first.

This post is the list of words I ended up matching, and the small traps behind them.

The label words

The shop's own unit price usually sits next to a label. These are the ones the extension reads today:

Language	Label
German	Grundpreis, Literpreis, Kilopreis
French	prix au kg, prix au litre, prix unitaire
Italian	prezzo al kg, prezzo al litro, prezzo unitario
Spanish	precio por kg, precio por litro, precio unitario
Portuguese	preço por kg, preço por litro
Dutch	prijs per kilo, prijs per liter, eenheidsprijs
Croatian / Bosnian	cijena za kg, cijena za j.m., jedinična cijena
Polish	cena jednostkowa, cena za kg
Czech	jednotková cena
Hungarian	egységár
Romanian	preț unitar, preț pe kg
Swedish	jämförpris
Norwegian / Danish	kilopris, literpris, enhedspris
Finnish	yksikköhinta, kilohinta

All of it is one case-insensitive regex anchored to the start of the label, so "Grundpreis" matches but a product called "Kilopreis-Knaller" in a title does not get treated as a label.

Piece counts

Eggs, rolls and coffee capsules are priced per piece. Each language has its own abbreviation, and most of them are two or three letters:

szt, sztuk (PL)  ks, kusů (CZ, SK)  db, darab (HU)  buc, bucăți (RO)
stuks, stuk (NL)  pz, pezzi (IT)  uds, unidades (ES)  un, unid (PT)
pièce, pièces (FR)  st, styck (SE)  stk, styk (DK, NO)  kpl (FI)
tk (EE)  vnt (LT)  gab (LV)  τεμ (GR)  бр (BG)  kom (HR, BA, RS)

Short words are dangerous. "db" is also the start of plenty of Polish words, "tk" shows up inside other text, and "1st" in English is an ordinal, not one Swedish piece ("st"). The rules that keep this safe: the unit must follow a number, must not be followed by another letter, and "1st" without a space is skipped. So "10 db" is ten pieces, while "dbałość" and "1st Grade Apples" are ignored.

The second rule: weight or volume beats a piece count. "Kawa kapsułki 10 szt. 52 g" is a 52 g pack, priced per kilo, not per capsule. People compare coffee by weight, and shrinkflation usually hides in the grams.

The third rule: when a product has no weight or volume at all, pack-count phrases such as "10er Pack", "12-pack", "12-p", "pack of 10", "lot de 12", "boîte de 10" or "confezione da 10" count as pieces. That is how Irish, German, Swedish, French and Italian shops usually title egg boxes.

Three decimals and decimal commas

The bug that cost me the most time was this one: Mercadona prints "1,367 €/L". My price regex expected at most two decimals, so it skipped the "1," and happily matched "367 €/L" in the middle of the figure. The cross-check then showed 367 euros per litre. The fix was to require that the number starts the figure (no digit or comma right before it) and to accept up to three decimals, because no supermarket sells milk at hundreds of euros a litre.

Why bother

France (since July 2024) and Hungary already make shops flag shrinkflation on the shelf. Everywhere else, the unit price is the only tool shoppers have, and it is often printed in small grey type. Reading it reliably in every language is the boring part that makes everything else work.

The code is MIT licensed and has no network requests of its own: it only reads the page you are looking at. If your shop prints its unit price with a word not in this list, open an issue with the product link and I will add it.

Website and install: https://emirmrkaljevic.github.io/shrinkwatch-eu/
GitHub repo: https://github.com/emirmrkaljevic/shrinkwatch-eu

This article was written with AI assistance. The code, tests and examples come from the repo.