How does the importer handle products (create/update by SKU) idempotently?

Products are imported idempotently by SKU (falling back to name): for each row the importer creates the product if its SKU is new, or updates the existing product if the SKU already exists, rather than blindly inserting duplicates. Work is done in batches (around 500 at a time) for performance on large catalogues. Idempotency means you can safely re-run an import — to apply corrections or…

Products are imported idempotently by SKU (falling back to name): for each row the importer creates the product if its SKU is new, or updates the existing product if the SKU already exists, rather than blindly inserting duplicates. Work is done in batches (around 500 at a time) for performance on large catalogues. Idempotency means you can safely re-run an import — to apply corrections or add new rows — without creating duplicate products or losing existing links; the same file run twice produces the same end state. This is the same idempotency principle the platform applies to billing and "mark paid" operations.