The search API

ids_set_encoding

is used to inform the Search Server of the encoding used by the client application for UNICODE columns (data type ’W’).

If an encoding has been specified that is different to the encoding used by the Search Server (UTF-16), the search data will be converted prior to searching and similarly, prior to the return of the result set.

UTF-16 UTF-8 conversions occur on the machine running the Search Server. UTF-16 data is assumed to be encoded in the byte order of the Search Server’s machine.

Oracle W fields are stored as UNICODE in the database’s national character set using

NCHAR/NVARCHAR2

data types. Upon retrieval by the Search Server the data is converted to UTF-16 (if necessary). If the caller’s search data is encoded differently, the caller must call

ids_set_encoding

to inform the Search Server.

IIR stores and retrieves data for W fields as UNICODE characters encoded as UTF-16. If the caller’s search data is encoded differently, the caller must call

ids_set_encoding

to inform the Search Server.

For example, if the client’s search data is encoded as UTF-8 the Search Server will convert incoming data from UTF-8 to UTF-16, perform a search and translate the search results back to UTF-8.