Almost every CSV import failure comes from a small set of mistakes. They are not exotic: a comma in the wrong place, a currency symbol in a price, a SKU typed twice. Learn these five and you will recognize them in any file — and understand why CSV Pilot exists to prevent them before they reach Hyperzod.
1. Shifted columns
A description like "soft, breathable cotton" contains a comma. Without quotes, the parser splits it into two cells and every value after it moves one column left. Prices land in descriptions, categories land in prices, and the import fails on random rows — or worse, succeeds with corrupted data.
Why it happens: hand-editing in a text editor, or a vendor file that was never validated.
Avoid it: quote any value containing a comma, quote or newline. CSV Pilot quotes descriptions automatically when it generates or exports a catalog, and its preview parses the file exactly as the importer will — if the preview looks right, the import will be right.
2. Numbers with extras
₹1,299.00 is not a number to a CSV parser — it is text with a comma in it. Currency symbols, thousand separators and locale decimal commas break price columns and can even shift columns like mistake #1.
Avoid it: keep price, inventory and SKU columns plain — 1299.00. Currency lives in the store settings, not the data. CSV Pilot writes numbers without symbols and validates them before export, so a "price" that is actually text is flagged while you can still fix it.
3. Duplicate SKUs
Two rows with the same SKU overwrite each other on import — one product silently replaces another. No error appears; a product just disappears. Duplicates creep in from copy-paste, from merging vendor files, and from reusing a code that "felt unique."
Avoid it: check uniqueness before import. CSV Pilot's validation flags duplicate SKUs as errors while you still can fix them — rename one, or merge the rows deliberately.
4. Broken variant cells
The VARIANTS cell packs name, price, cost, inventory, description, image for every choice into one field. One missing comma or unquoted description shifts every field after it, and the variant arrives with the wrong price — or no price at all. Because several variants share one cell, one broken record can hide behind two correct ones.
Avoid it: never hand-edit variant cells. Generate them from a description ("sizes Small, Medium, Large") or convert them from a Shopify or DeliverLogic file, and let CSV Pilot write the cells with proper quoting and field counts.
5. Wrong encoding
café becomes café when a file is saved in the wrong encoding. Non-English product names and descriptions suffer first, and the damage is invisible in some editors and obvious in others.
Avoid it: use UTF-8 everywhere. Excel's plain "CSV" save option is the usual culprit — its "CSV UTF-8" option is the safe one. CSV Pilot's exports are UTF-8, and its preview shows you exactly how the file will be read: if the preview looks right, the import will too.
The pattern behind all five
Every mistake above happens when a person edits raw CSV text by hand. The errors are structural — quoting, field counts, encodings — and structure is exactly what humans are bad at checking row by row.
The structured approach is the opposite: describe the data, let the tool encode it, and review the result in a preview that parses the file exactly as the importer will. Validation flags what the eye misses — duplicates, missing fields, malformed numbers. That is the CSV Pilot workflow, and the reason its exports pass validation before they reach Hyperzod.