Most failed imports are caught too late — after the file has been uploaded, when the platform reports an error on row 400. At that point you fix the row, re-export, re-import, and hope the next error is on a row you already checked. Validating the CSV before import catches the same problems in seconds, while you can still fix them. This guide explains what validation checks, why each check matters, and how CSV Pilot runs them for you.
What can go wrong in a product CSV
- Missing required fields — a product without a name, selling price, cost price or category fails the import or arrives as garbage data.
- Duplicate SKUs — two rows with the same SKU (case-insensitively) overwrite each other. One product silently replaces another; no error tells you which.
- Price problems — negative prices, prices with currency symbols, or compare prices lower than selling prices.
- Broken variants — a VARIANTS cell with a missing field shifts every value after it, so a variant arrives with the wrong price or no image.
- Nested-option violations — a MULTIPLE group containing nested option groups, unbalanced braces, or nesting beyond the confirmed depth.
- Formula injection — values starting with
=,+,-or@can execute as formulas when the exported file is opened in a spreadsheet.
Each of these is cheap to fix before an import and expensive after.
What validation checks
CSV Pilot validates products against the Hyperzod import requirements:
- Every product has a name, selling price, cost price and category
- Prices are plain numbers, and compare prices are higher than selling prices
- SKUs are unique case-insensitively
- Tax percent is a non-negative number and status is active or inactive
- Option groups have valid variants with valid prices; MULTIPLE groups cannot nest
- The export is guarded against formula injection
The result appears as an error and warning count next to the preview. Errors are things that must be fixed — the import will fail or corrupt data. Warnings are things worth checking — usually inferred values or unusual numbers.
How to fix flagged issues
Validation results update live as you work:
- Look at the error count in the header.
- Click through the flagged products in the preview.
- Fix each cell directly — the error disappears when the value is valid.
- Export when the count reaches zero.
Because validation runs on every change, you never work long with broken data. The feedback loop is immediate: change a price, see the validation update.
The review-before-export habit
Make validation part of your workflow, not a final step:
- After generating a catalog, check inferred fields (highlighted in the preview).
- After converting a Shopify or DeliverLogic file, spot-check three products: simple, with variants, with add-ons.
- After editing, re-check the products you changed.
- Before every export, confirm the error count is zero.
Why validation beats trial-and-error importing
The trial-and-error approach — import, read the error, fix the row, re-import — works for a 10-row file. At 500 rows it becomes a loop of failures, each costing an import cycle and a cleanup. Validation moves the same checks to before the import, where fixing a row takes seconds instead of a re-import cycle.
A CSV that passes validation is not guaranteed to match every destination-specific business rule. It has passed CSV Pilot's structural checks, which address common causes of import rejection, but the destination platform may enforce additional rules. When the error count is zero, review the warnings and a small test import before replacing a live catalog.