CSV structure and character encoding are separate
Correct comma-separated columns do not determine how bytes become characters. Reading UTF-8 bytes with another character set can turn Korean into question marks or garbled text.
Why UTF-8 can still behave differently
Some programs use a BOM or the operating system default to guess encoding. The same CSV can therefore open differently in a browser, text editor and spreadsheet app.
Do not overwrite the original when text is broken
Saving a garbled file can alter the original byte data and make recovery harder. Work on a copy and try the appropriate encoding first.
Consider the recipient’s environment
UTF-8 is common for web data exchange, but some business software requires another encoding. Follow the import rules of the destination system when they are specified.