This guide describes the principles of CSV-based media management in AI Commerce . You work in a table (export → edit → import), which allows you to update media categories and media files (e.g. PDF) attached to products at once without having to open individual products. The method is fast, repeatable and auditable.

Workflow (export → edit → import)

1. Create an export template: use Category CSV Export or File CSV Export.
2. Edit values in the table: make the necessary changes directly in the CSV.
3. Import changes: use Category/File CSV Import.
4. Keep the same column structure as in the export.

Multilingual columns

Language-specific fields are marked with square brackets, e.g. name[fi] , description[en] . This keeps the descriptions consistent across languages.

Filenames and automatic retrieval

• Only filename field is interpreted during import.
• You can reference a file previously uploaded to uploads folder by name alone or by providing a full external URL; the file will be retrieved and saved.
• url is for viewing and does not affect the import.
• When editing only metadata (e.g. description or order), leave filename blank - the existing file reference will be preserved.

File storage and scalability

Files are stored in Amazon S3. S3 does not act as a traditional folder disk, so the size or number of files in a single “folder” does not constitute a bottleneck.

Auditing

CSV runs and changes are traceable using CSV history and history views, which supports error location and auditability.

Particularly suitable for

• Mass repairs
• Renaming
• Unified media sorting without product-specific clicking