Document Uploads
Upload product data files and let MerchantOps AI extract structured product information automatically.
Learn how to upload and process product data files.
Supported File Formats
MerchantOps can process the following file types:
| Format | Extensions | Best For |
|---|---|---|
.pdf | Product catalogs, spec sheets, brochures | |
| Excel | .xlsx, .xls | Vendor price lists, inventory exports |
| CSV | .csv | System exports, data feeds |
Tip: For best results with PDFs, use files that contain tabular data or structured product listings. Our AI works best with consistent, well-formatted documents.
Uploading a Document
To upload a new document:
- Navigate to Lakehouse → Documents in the sidebar
- Click the "Upload Document" button
- Either drag and drop your file or click to browse
- Select the file from your computer
- Click "Upload" to start processing

Upload Options
When uploading, you can optionally specify:
- Technology/Brand: Associate the document with a specific brand source for better organization
- Product Type: Pre-select the product type to apply to extracted products
- Notes: Add internal notes about the document source
Document Processing
After uploading, MerchantOps processes your document through several stages:
| Status | Description |
|---|---|
uploading | File is being uploaded to the server |
processing | AI is extracting product data from the document |
mapping_required | Column mappings need review before proceeding |
completed | Products successfully extracted and ready for review |
failed | Processing encountered an error |

Viewing Extraction Results
Once processing completes, you can view detailed results:
- Click on the document in the Documents list
- The document detail page shows extraction statistics
- View the list of extracted products
- See any errors or skipped rows
Extraction Statistics
Each processed document displays:
- Total Rows: Number of rows detected in the document
- Products Created: Rows successfully converted to products
- Skipped Rows: Rows that couldn't be processed
- Success Rate: Percentage of successful extractions

Handling Errors
Common Issues
| Issue | Cause | Solution |
|---|---|---|
| Low success rate | Inconsistent data formatting | Review column mappings, clean source data |
| Missing fields | Columns not mapped correctly | Update mapping configuration |
| PDF not parsing | Scanned image or complex layout | Try OCR-optimized version or Excel export |
| Encoding errors | Non-UTF8 characters in CSV | Save CSV with UTF-8 encoding |
Skipped Rows
Rows may be skipped for various reasons. Click on "View Skipped Rows" to see:
- Which rows were skipped
- The reason for each skip
- The original row data
Common skip reasons include:
- Missing required fields (like product name or SKU)
- Invalid data format in key fields
- Duplicate product keys
- Empty rows
Reprocessing Documents
If you update column mappings or want to re-extract data, you can reprocess a document:
- Open the document detail page
- Click the "Reprocess" button
- Confirm the action
Note: Reprocessing will create new lakehouse products. Existing products from the previous extraction are not automatically deleted.
Bulk Uploads
For large batches, you can upload multiple documents at once. Each document is processed independently in its own background job.
When uploading multiple files:
- Select multiple files in the file picker
- Or drag and drop multiple files into the upload area
- Each file creates a separate document entry
- Monitor progress on the Jobs page
Best Practices
File Preparation
- Clean your data: Remove empty rows and formatting before uploading
- Use consistent headers: Column names should be clear and match across files
- Include unique identifiers: Always have a SKU or product key column
- UTF-8 encoding: Save CSVs with UTF-8 to avoid character issues
Organization
- Name files descriptively: Include vendor name and date in filenames
- Associate with Technologies: Link documents to their brand source
- Process similar files together: Batch uploads from the same vendor
Next Steps
After uploading your documents:
- Review column mappings if prompted
- Check extracted products for accuracy
- Monitor job progress for large uploads