How to work with Unicode in Oracle SQL Developer?

Question

Oracle

There is a download in csv format. If you open it in Notepad++, then the encoding will be written as UTF-8-BOM. I import data from a file into a separate table in the database using Oracle SQL Developer and specify the UTF-8 encoding type.
As a result, after importing, I see that some of the special characters (such as fractions ½ and ¼ with one character, exotic currency signs such as the ruble symbol, etc.) have been replaced with a question mark (?). If you export this table again to csv, then the resulting file will also have question marks, and not the original characters. How to make sure that when importing and then exporting, the data is saved without distortion and all characters remain in their original form?