Multilingual Problem in XMLTYPE

Question

Hi Experts, 
My database as follows:
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
PL/SQL Release 11.2.0.3.0 - Production
CORE 11.2.0.3.0 Production
TNS for Linux: Version 11.2.0.3.0 - Production
NLSRTL Version 11.2.0.3.0 - Production
NLS_CHARACTERSET WE8ISO8859P1
Im storing XML in one of the table in my database in XMLTYPE column. But when my XML node contains some multilingual value( Ex: Chinese) it is showing junk value. Is there any way to store as it is?
INSERT INTO web_xml_file  (file_no, web_file_data, record_status, user_created, created)   VALUES (v_file_no, xmltype.createxml(p_in_xml), 'UNP', USER, SYSDATE);
  FILE_NO       NUMBER,
  WEB_FILE_DATA XMLTYPE,
  RECORD_STATUS VARCHAR2(3),
  ERROR_MSG     VARCHAR2(4000),
  USER_CREATED  VARCHAR2(50),
  CREATED       DATE,
  USER_CHANGED  VARCHAR2(50),
  CHANGED       DATE

Answer 1

The following seems to work but be aware it's still a hack, and certainly not bulletproof.
The best way to handle this is to have the db charset migrated to AL32UTF8, or to do the parsing externally.
select utl_i18n.raw_to_nchar(utl_i18n.string_to_raw(skywardsnumber), 'AL32UTF8') as skywardsnumber
     , utl_i18n.raw_to_nchar(utl_i18n.string_to_raw(title), 'AL32UTF8') as title
     , utl_i18n.raw_to_nchar(utl_i18n.string_to_raw(firstname), 'AL32UTF8') as firstname
     , utl_i18n.raw_to_nchar(utl_i18n.string_to_raw(familyname), 'AL32UTF8') as familyname
from xmltable('/data/passengerinformation'
       passing xmltype(dbms_xslprocessor.read2clob('XML_DIR','multi.xml', nls_charset_id('CHAR_CS')))
       columns skywardsnumber  varchar2(100) path 'skywardsnumber'
             , title           varchar2(100) path 'title'
             , firstname       varchar2(100) path 'firstname'
             , familyname      varchar2(100) path 'familyname'
     ) x ;
Explanations :
1) I've put your sample XML in a file saved with UTF-8 encoding, and shipped it to a database directory (DOC_DIR).
2) I access the file as CLOB using DBMS_XSLPROCESSOR.read2clob, and (here's the trick) considering the file in the db charset (mine was WE8ISO8859P15). As a consequence, all multibyte characters are converted to single byte characters in the DB charset.
3) I use the XMLType constructor to build an XMLType instance out of the CLOB and parse it with XMLTABLE.
4) That gives me four columns that I then convert to byte sequences (utl_i18n.string_to_raw), and finally to NCHAR (utl_i18n.raw_to_nchar) by specifying the original encoding of the file.