Product Feed Validator: Prepare, Validate and Compare CSV or Excel Files
A useful feed validator should tell you what it checked, identify the affected product and explain the next action. It should also refuse a file it cannot interpret safely.
Prepare a readable input
Place one header row on the first row and at least one product beneath it. Use an UTF-8 CSV, optionally with a byte-order mark, or an XLSX file with exactly one populated worksheet. Comma, semicolon and tab separators are supported for CSV. Quote fields containing the delimiter or a line break rather than deleting punctuation from product descriptions.
Use attributes such as id, title, description, link, image_link, price and availability. Recognized aliases can be proposed as header changes, but two headers that map to one attribute are rejected as ambiguous. Unknown columns remain in the parsed data and exports; they are not a promise that their business rules have been validated.
Preserve text that looks numeric
Store SKU 00011 and GTIN 036000291452 as text before exporting. A spreadsheet may otherwise remove leading zeroes or convert a long identifier to scientific notation. FeedDoctor preserves text values and certain explicit zero masks in XLSX, but cannot recover precision that another program already discarded. Reopening the source and output as text is more useful than changing cell appearance after the fact.
Understand the three finding levels
- Error: a file-level or product-data problem found by the baseline checks, such as a missing title, duplicate ID or malformed price.
- Warning: a condition needing context or manual review, such as missing brand information or a multiple-GTIN field.
- Info: an explanation or optional normalization proposal. It does not mean a product is approved.
Products without errors may still have warnings. The same product can appear in both the error and warning groups. Total findings are therefore a different measure from the number of affected products. Use the filters and search field to investigate a specific ID, field or message.
Try a small reproducible example
id,title,description,link,image_link,price,availability,gtin,brand
00011,Ceramic cup,Blue ceramic cup,https://example.com/cup,https://example.com/cup.jpg,19.95 USD,in_stock,036000291452,ExampleThis example uses placeholder URLs and a syntactically valid barcode to illustrate the file structure. It is not a real product listing or evidence of barcode ownership. FeedDoctor checks URL syntax rather than fetching the image or page. Use the built-in sample analysis for a real run of the engine and compare its report with the example panel.
Work within bounded processing
The current public service accepts up to 5,000 products and 10 MiB per upload. These are operational limits, not a paid tier. A 10,000-row file is rejected before normal analysis rather than silently truncated. Large cells, excessive columns and unsafe workbook archives have separate safeguards. The service processes one analysis at a time on the current server and asks another uploader to retry when busy.
Know the supported format boundary
FeedDoctor 0.4.0 accepts supported Merchant RSS/XML and native Shopify/WooCommerce CSV exports through dedicated, bounded adapters. Use the focused tools to inspect mapping and supported structures. A verified currency and store URL can supply context for native files; missing prices, stock facts and variant destinations are never invented. Unsupported schemas are rejected. The original remains available, and any converted Merchant export still needs destination-specific review.
Compare and retain your results
Preview safe changes, apply only selected proposals and recheck the result. Download the original, the corrected feed and a detailed CSV or XLSX report. Files expire after the configured temporary retention period and are not permanent account storage. Keep your own authorized copy if you need to continue later.
References and next steps
These instructions describe the current FeedDoctor implementation. Check the official requirements for your product, target country and destination before submitting data.