Change CSV separator

Comma, semicolon, tab and pipe, on both sides. The to option picks the new one and defaults to semicolon. The from option can be left blank, which is the default, and then the current separator is detected. There is no support for arbitrary characters or multi-character separators.

It parses the first complete record with each of the four candidates and keeps whichever produces the most fields. It works on the first record, not the first physical line, so a header with a quoted field containing a newline is handled correctly. A header that genuinely contains every candidate character inside quoted fields can still confuse it, so set from yourself if the guess is wrong.

Because the detected or supplied from value matched the to value, so there would be nothing to change. The message names the separator, spelling tab out in words. Pick a different target, or set from explicitly if you think the detection was wrong.

Yes, and re-quoted for the new separator. A cell that needed quoting for a comma might not need it for a tab, and a cell that was plain might now need quotes because it contains the new separator. The writer works that out per cell, so a value containing a semicolon is safe in a semicolon separated file.

They are normalised. The parser accepts CRLF and bare LF on the way in, and the writer always emits LF, with a trailing newline at the end of the file. If you need CRLF back, run Line-ending and BOM normalizer on the result.

Your file name with -resep.csv on the end, and the result line reports the change, for example from comma to semicolon. The file is converted in your browser with no upload, and the whole thing is held in memory while it runs.