Dataset documentation
Everything you need to import The Complete Archive and work with its product records, ingredients, nutrition and images.
Choose your format
JSONL
Read one product per line, with nutrition sets and nutrients nested inside each record.
CSV
Import three related tables: products, nutrition sets and nutrients. Arrays and objects are represented as JSON text.
SQLite
Open the dataset as a single database, with related tables, indexes and constraints ready to query.
Work with barcodes
Store barcodes as text to preserve leading zeros. The dataset uses 14-digit GTINs; normalise supported scanned barcodes to the same format before matching.
Join nutrition records
Join nutrition sets to products using product_id. Join nutrient rows to their nutrition set using both product_id and nutrition_set_id.
Connect product images
Match a product’s image_name to the corresponding image filename. Use images-manifest.json to identify the ZIP archive containing that image.
Interpret values correctly
Keep missing values distinct from recorded zero values. The field reference explains each field’s type, meaning and format-specific representation.
Start with the sample
Use the free sample to test your import process before loading the full dataset. It includes JSONL, CSV and SQLite files, schema documentation and referenced product photos.