How CSV row repair works
Upload a CSV with comma, semicolon, tab or pipe delimiters. The parser respects quoted separators and embedded line breaks. It determines the widest record, adds empty cells only to shorter rows and writes a UTF-8 comma-separated result. Every original cell is compared with the output after parsing it again.
What repair cannot recover
Missing source values cannot be guessed. Extra cells are retained as separate columns rather than merged into a neighboring cell; review the resulting column headers before importing. Unclosed quotes, invalid UTF-8, NUL bytes and spreadsheet-formula injection risks cause a safe rejection instead of an unreliable download.
Before importing the repaired file
Review the number of columns, decimal separators, date formats and any empty padded fields. The browser converter supports up to 10 MB, 50,000 rows, 200 columns and 100,000 characters per cell. For converting a consistent CSV into records, use CSV to JSON; to export records back to a spreadsheet, use JSON to CSV.
Why ambiguous separators are not repaired automatically
A file such as id,name;role with matching mixed separators on every row can describe more than one valid table. Automatically choosing comma or semicolon would silently change which values belong together. In that case the tool stops with an ambiguity message; open the original in a text editor and normalize the delimiter before repairing. A separator inside a correctly quoted cell does not create this ambiguity.
Dedicated repair versus batch preservation
This converter deliberately writes a new comma-separated UTF-8 file with CRLF record endings; it does not promise byte-identical source formatting. If the original uses mixed CRLF, LF and CR endings and you need to retain those record boundaries while padding short rows, choose universal batch verification instead. Review the resulting cells in either workflow before importing them into a spreadsheet.