Sort of along the lines of "education"... I've been helping a colleague who is using the Oracle database and trying to work through a customer's character conversion and mojibake issues. I started suspecting the NLS_LANG variable and looked up some references, and found the following alternative facts on the Oracle FAQ and community pages:
> SQL> SELECT DUMP(col,1016)FROM table; > > Typ=1 Len=39 CharacterSet=UTF8: 227,131,143,227,131,170 > > returns the value of a column consisting of 3 Japanese characters in > UTF8 encoding . For example the 1st char is 227(*255)+131. and: > While UTF8 uses only 2 bytes to store data AL32UTF8 uses 2 or 4 bytes. Unicode and UTF-8 have been around a long time by now. The fact that there is still fake news like this out there, steering our less Unicode-aware colleagues waaay down the wrong path, is disconcerting. -- Doug Ewell | Thornton, CO, US | ewellic.org
