How do I add or update food labels for many products with a CSV file?
On this page
Use a CSV file when you label many products at once, or when your product data already lives in a spreadsheet. You can import up to 1,000 rows with label data per file; rows without label data don’t count and are skipped.
Steps
- In the app, open Food labels and click Import CSV.
- Click Download template (in the import window, or at the top of Food labels). It has one row per product (and one row per variant for products labelled per variant) and one column per field. Export CSV at the top of Food labels gives the same format with your current labels.
- Fill in the file in your spreadsheet and save it as CSV (UTF-8). Commas or semicolons as separators, decimal commas (
0,5) and line breaks inside a cell (for example in ingredients) are all fine. - In the import window, choose how rows are matched (Match rows to products by) and how a row changes the label, then drop the CSV file on the drop zone or click it to select the file. Nothing is saved yet: you see a preview.
- Check the preview, choose Save as drafts or Publish the rows without errors under What should the import do?, and confirm with Save N drafts or Publish N labels.
Matching rows to products
- Product ID (
product_idcolumn): the default, and the safest. If a row also has ahandle, it must be the handle of that product: a file exported from another store is stopped on every line instead of writing to the wrong products. - Handle (
handlecolumn): useful for a file made in another store (for example a staging store), where product IDs are different. Theproduct_idcolumn is ignored. Handles are compared without regard to upper or lower case. - SKU (
skucolumn): only a SKU that belongs to exactly one variant in your store is used. The row writes that variant’s product label, or the variant’s own label when the row fills invariant_override(see variant labels).- An empty SKU, or a SKU no variant has, stops that row.
- If several variants share the SKU, the row is stopped and the error lists each one (product ID/variant ID). The app never picks one for you. Fill in
variant_idorproduct_idon that line to choose, or match the file by product ID or handle. - When the SKU belongs to a single variant, the row’s
product_idis not used (like the handle match, so a file from another store still imports), but a filled-invariant_idmust be that variant.
The export and the template have a sku column: the variant’s SKU on variant rows, and the SKU of products that have a single variant.
Replace label or update fields
- Replace label (default): each row is the whole label. An empty cell empties that field.
- Update fields: only the cells you fill in change the label. An empty cell, or a column that isn’t in the file at all, keeps the current value. Your file can be as small as
product_id,salt.
To clear a field in Update fields mode, write [clear] in the cell (upper or lower case, spaces around it are ignored). In Replace label mode [clear] is the same as an empty cell.
0 and an empty cell are different: 0 is a value (for example “0 g salt”), an empty cell is no value. A list cell (allergens, contains, may_contain, diet) replaces the whole list; it doesn’t add to it.
In Update fields mode, the row is checked together with the label it updates: if the result isn’t a valid label (for example a nutrition table with only one value), the row is an error.
The preview
- The badges count the rows ready to import, the rows with errors, the unchanged rows and the rows without label data. A sentence tells how many product labels and variant labels the import changes.
- Changes per row lists each row; open one to see every field that changes, its current value next to the value after the import.
- Rows with errors are listed with the reason and are never written, whichever option you choose.
- When the import replaces existing labels, the preview offers Download current labels: a CSV file of those labels as they are now. Importing that file with Replace label and the same draft or publish choice puts them back.
Exporting your labels and importing the file again changes nothing.
Large files, interruptions and retries
The import writes one row after another from your browser tab, so keep the tab open until it finishes; 1,000 rows take a few minutes. When Shopify asks the app to slow down, it waits as long as Shopify says and continues.
If the connection drops or Shopify stops answering, the import stops at that line. The result lists the lines that were written, the lines that weren’t, and Continue from line N: it checks the file again (against the labels as they are now) without the lines already written. Importing the whole file again later is also safe: rows that were written are unchanged and skipped.
A line whose label was changed after the file was checked (for example in the editor in another tab) is not written. Click Check the file again to see a new preview.
Limits
- 1,000 rows with label data per file. Split a larger catalogue into several files.
- The import runs in your browser: closing the tab stops it. Continue it with the same file.
- The
titlecolumn only helps you recognise products and is never imported. - Column names must be the import’s own (as in the template). You can also map a supplier’s own column names instead.
- Your products already describe their ingredients? Create label drafts from product descriptions or metafields instead of filling in a file.