How does WTX (websphere transformation extender) know whether the character is ASCII or UNICODE?

ASCII is 8-bit value. Unic开发者_Go百科ode may be 8 or 16 or 32 bit value. If I define subclass as character how does WTX know whether it is 8 or 16 bit character?

Setting an item's subclass to character is only one half of the solution. You also have to set the language (defaults to "Western") and, more important, the character set. If you choose UTF-8 (-16, -32), the parser is capable of recognizing multi-byte characters and will read them properly (given that the document being parsed is encoded in the type tree's encoding, of course).