Help & User Guide

Regional Delimiters

CSV files use different characters to separate columns depending on the region they were created in (e.g., North America usually uses commas, while Europe often uses semicolons). The CSV Rescue Kit automatically attempts to detect the correct input delimiter.

If your columns look incorrect after loading a file, you can manually override the input delimiter:

• Comma ( , ) - Standard globally.
• Semicolon ( ; ) - Common in European regions where the comma is used as a decimal separator.
• Tab - Used in TSV exports.

You can also use the Output Delimiter setting to convert a file from one regional format to another (e.g., upload a semicolon CSV and export it as a comma CSV).

Browser Limitations & File Size

Because the CSV Rescue Kit runs entirely offline inside your web browser (ensuring complete privacy for your sensitive data), its processing capability is bound by your device's memory.

• Recommended Size: Up to 50MB. Modern browsers can comfortably parse and clean files in this range within seconds.
• Maximum Practical Size: ~150MB to 200MB. Depending on your system RAM, attempting to load files larger than this may cause your browser tab to freeze or crash because the raw text and the structured data matrix must both be held in memory simultaneously.

Tip for massive files: If you have a 1GB+ CSV file, we recommend splitting it into smaller chunks before processing them through the application.

Edge Cases & Corrupted Files

While the parsing engine is robust and strictly follows RFC 4180 standards for handling quoted fields and escaped line breaks, some malformed files might yield unexpected results:

• Mixed Delimiters: If a system erroneously generated a file that mixes semicolons on some lines and commas on others, the auto-detector will choose the most frequent one. The unselected characters will be treated as raw text within the column, potentially skewing the column count.
• Unclosed Quotes: If a field begins with a double quote " but the file truncates or fails to close the quote before the end of the line, the parser may absorb subsequent lines into a single massive cell until it finds a closing quote.
• Corrupt Line Endings: Mac (CR), Unix (LF), and Windows (CRLF) line endings are all normalized during parsing. However, completely stripped line breaks will result in a single, continuous row.