in DuckDBPyRelation.to_csv, use of any quoting value other than csv.QUOTE_ALL throws an exception.

This is annoying because the fairly common csv.QUOTE_STRINGS is pretty useful for differentiating between strings and numbers in CSV.

The underlying CSV writer doesn't support all the usual options either but there is a FIXME

 // FIXME: we could gain some performance here by checking for certain types if they ever require quotes
 // (e.g. integers only require quotes if the delimiter is a number, decimals only require quotes if the
 // delimiter is a number or "." character)

... which while it doesn't explicitly mention compatibility with the other common CSV options is at least pointing in a similar direction.