
On 09/25/2016 04:58 PM, Yves Dorfsman wrote:
On 2016-09-25 03:39, Dennis Pallett wrote:
Found the answer myself on https://www.monetdb.org/wiki/MonetDB_type_system
Character \200 works to indicate a real NULL in MonetDB.
Dennis: Have you tried it? Is this 200 decimal (so the "È" character?) or 200 octal (the "€" sign?)?
The wiki says it's for char, varchar, text etc... Can we not use unicode for text?
It's neither. MonetDB is fully UTF-8 internally, and so the text in a binary import also has to be UTF-8. In UTF-8, the \200 byte (which incidentally is standard C notation for a byte with octal value 200, i.e. decimal 128) is an illegal value. UTF-8 is a variable length encoding, but there is no character that is encoded with a byte that starts with \200. -- Sjoerd Mullender