How to support non alphanumeric characters when using WORLD_LEXER?

Question

BASIC_LEXER has an attribute of printjoins which we can specify the non alphanumberic characters as normal alphanumberic in query and included with the token. However, WORLD_LEXER doesn't have this attribute. So in order to use some non alphanumberic characters and treat them as alphanumberic characters in Oracle Text Index, such as ><$, what should I can?
Thanks in advance for any help.

Answer 1

I use WORLD_LEXER to create Oracle Text Index to support UTF-8.
Below is the script to create table and index:
REM Create table
CREATE TABLE my_test
( id VARCHAR2(32 BYTE) NOT NULL,
code VARCHAR2(100 BYTE) NOT NULL,
     CONSTRAINT "my_test_pk" PRIMARY KEY ("id"));
REM create index
exec ctx_ddl.create_preference('stars_lexer','WORLD_LEXER');
exec ctx_ddl.create_preference('stars_wordlist', 'BASIC_WORDLIST');
exec ctx_ddl.set_attribute('stars_wordlist','substring_index','TRUE');
exec ctx_ddl.set_attribute('stars_wordlist','PREFIX_INDEX','TRUE');
-- create index for Table corrosion level
CREATE INDEX my_test_index
ON my_test(code)
INDEXTYPE IS CTXSYS.CONTEXT
PARAMETERS ('LEXER stars_lexer STOPLIST stars_stop WORDLIST stars_wordlist SYNC(EVERY "SYSDATE+5/1440") TRANSACTIONAL');
INSERT INTO my_test('1', 'English word1');
INSERT INTO my_test('2', '违违');
INSERT INTO my_test('3', '违违&^|违违');
When I query:
select * from corrosion_levels r where contains(r.CORROSION_LEVEL, '{%违${|违%}') > 0
ID CODE
3 违违$(|违违
2 违违
Actually, the result what I want is: 3 违违$(|违违
So the requirement is that all non-alphanumeric characters should be treated as normal alphanumeric charcters. Please tell me how to implement it.