A CSV file can open perfectly in a spreadsheet and still fail during import. Spreadsheet software often hides quoting, delimiter, encoding, and line-ending details that another system treats as strict input rules. A preflight check is simply a repeatable way to inspect those details before you upload.
1. Keep an untouched copy of the original
Before changing anything, duplicate the source file. A CSV repair should be reversible. If a conversion changes quoting, line endings, delimiters, or encoding, the original gives you a reliable reference.
2. Confirm the destination format
Do not treat every .csv file as interchangeable. The destination may require a particular header row, field order, encoding, delimiter, maximum file size, identifier format, or product relationship. Generic CSV validation can prove that the table is structurally readable, but platform rules are a second layer.
3. Validate UTF-8 when the destination requires it
Unexpected characters such as José or replacement symbols are usually encoding symptoms. If the destination requires UTF-8, first verify whether the byte stream is valid UTF-8 instead of blindly decoding it as another character set.
4. Detect the delimiter
A file can be named CSV while using semicolons, tabs, or pipes. Regional spreadsheet settings commonly produce semicolon-delimited files. If the importer expects commas, every row may appear to have one giant column.
5. Compare each row with the expected column count
If the header resolves to 12 fields, later rows should normally resolve to 12 fields too. Too few or too many fields often point to a stray delimiter or a quoting problem.
Find column-count mismatches →
6. Inspect quoted fields
Fields containing delimiters, line breaks, or quotes need correct CSV quoting. A single missing closing quote can make a parser consume several later lines as one field. A stray quote can shift the rest of a row.
7. Check headers for invisible mistakes
Headers can look correct while containing leading spaces, trailing spaces, duplicate names, or an unexpected BOM. Destination-specific imports may also require exact names. Treat headers as data that deserves its own validation step.
8. Review blank rows and duplicate records
Blank trailing lines are usually harmless, but blank rows in the middle of a file can surprise brittle importers. Duplicate records are more complicated: they may be valid repeats, accidental duplication, or separate variants. Never remove them automatically unless the destination rule is clear.
9. Look for spreadsheet side effects
Spreadsheet programs may convert long numbers to scientific notation, strip leading zeros, reinterpret date-like strings, or evaluate cells beginning with =, +, -, or @. These are not purely CSV syntax errors, so a safe repair tool should warn rather than guess the intended value.
10. Test a small import before the production import
Even a structurally perfect CSV can fail a business rule that depends on the destination account. Test a small representative file first. For a large commerce import, verify products, identifiers, images, variants, and update behavior before importing the full dataset.
What a preflight check cannot prove
An offline checker cannot know whether a SKU already exists in your store, whether a referenced product exists, whether your account has reached a platform limit, or whether a business rule changed after the checker was last updated. CSVFixLab therefore separates local structural checks from platform-specific guidance.