Suppliers send updated product feeds regularly — weekly, monthly, or even daily. FeedPrep uses a versioning system to keep track of every file you receive, so you never lose historical data and can always compare what changed.
How Versioning Works
When you upload a new file for an existing feed (from the same supplier), FeedPrep creates a new version of that feed rather than overwriting the previous data. This applies to manual uploads, URL re-fetches, and FTP/SFTP scheduled imports alike.
Each version is linked through a parent/child relationship:
- The original upload is the first version (the parent).
- Each subsequent upload becomes a new child version, linked to the same feed.
- The most recent version is treated as the active version for normalization and export.
Raw Data Preservation
Every version preserves its own raw data exactly as it was received. FeedPrep stores the original parsed content for each version independently. This means that even after uploading a newer version, you can go back and inspect the exact data that was in the previous file.
This is useful when you need to:
- Verify what a supplier sent at a specific point in time
- Investigate when a product was added or removed from a feed
- Debug normalization issues by reviewing the raw input data
Comparing Versions
FeedPrep allows you to compare changes between versions. This helps you understand what changed in a supplier's feed from one delivery to the next, including:
- New rows (products added by the supplier)
- Removed rows (products no longer in the feed)
- Changed values (price updates, description changes, attribute modifications)
Version comparison gives you confidence that your normalization rules are handling the new data correctly, and helps you spot unexpected changes early.
Accessing Previous Versions
All previous versions of a feed remain accessible from the feed detail page. You can browse the version history, open any past version to review its raw data, and compare any two versions side by side.
Previous versions are never deleted automatically. They remain available for as long as the feed exists in your workspace.
Versioning and Normalization
When a new version is created, your existing normalization configuration (column mappings, value mappings, transform rules) is automatically applied to the new data. You do not need to reconfigure anything — the adapter remembers your rules and applies them to each new version.
If the new version contains values that were not present in previous versions (for example, a new color or material), those will appear as unmapped values for you to review.