Ignore spelling differences?

Question

If I do a search for ...contains(column, 'catalogue'), I get 768 results. If I do a search for ...contains(column,'catalog'), I get 349 results.
How can I make Oracle Text ignore such spelling differences? (I hoped normal stemming would do it, but not in this case it doesn't!)
I know I can create a thesaurus and then add in a synonym, but (I believe!) I then have to change the query so that it's something like ...contains(column, SYN(catalog,default)), but that involves changing the query.
Is there a way of getting Oracle Text to ordinarily ignore this sort of American/UK spelling difference without having to alter the query that's being issued?

Answer 1

Roger's suggestion is a good one. However, note that if you do that, you will only be able to search for the one you choose to convert to. For example, if you replace all jail with gaol, you will then only get results searching for gaol, not jail, but it will return the original unchanged values. It depends on what you want. Please see the brief demo below. I used simple replace statements, but you could use regular expressions.
SCOTT@orcl_11g> create table test_tab (id number, test_col varchar2 (60))
  2  /
Table created.
SCOTT@orcl_11g> insert all
  2  into test_tab values (1, 'mail catalog')
  3  into test_tab values (2, 'postal catalogue')
  4  into test_tab values (3, 'male cat on a log')
  5  into test_tab values (4, 'mail')
  6  into test_tab values (5, 'catalog')
  7  into test_tab values (6, 'jail')
  8  into test_tab values (7, 'gaol')
  9  select * from dual
10  /
7 rows created.
SCOTT@orcl_11g> CREATE OR REPLACE PROCEDURE your_proc
  2    (p_rowid IN           ROWID,
  3       p_clob     IN OUT NOCOPY CLOB)
  4  AS
  5    v_col                test_tab.test_col%TYPE;
  6  BEGIN
  7    SELECT ' ' || test_col || ' '
  8    INTO   v_col
  9    FROM   test_tab
10    WHERE  ROWID = p_rowid;
11    v_col := REPLACE (v_col, ' catalog ', ' catalogue ');
12    v_col := REPLACE (v_col, ' jail ', ' gaol ');
13    -- and so on for each replacement
14    v_col := TRIM (v_col);
15    DBMS_LOB.WRITEAPPEND (p_clob, LENGTH (v_col), v_col);
16  END your_proc;
17  /
Procedure created.
SCOTT@orcl_11g> SHOW ERRORS
No errors.
SCOTT@orcl_11g> BEGIN
  2    CTX_DDL.CREATE_PREFERENCE ('your_datastore', 'USER_DATASTORE');
  3    CTX_DDL.SET_ATTRIBUTE ('your_datastore', 'PROCEDURE', 'your_proc');
  4  END;
  5  /
PL/SQL procedure successfully completed.
SCOTT@orcl_11g> create index test_idx on test_tab (test_col)
  2  indextype is ctxsys.context
  3  parameters ('DATASTORE your_datastore')
  4  /
Index created.
SCOTT@orcl_11g> select token_text from dr$test_idx$i
  2  /
TOKEN_TEXT
CAT
CATALOGUE
GAOL
LOG
MAIL
MALE
POSTAL
7 rows selected.
SCOTT@orcl_11g> select * from test_tab where contains (test_col, 'catalog') > 0
  2  /
no rows selected
SCOTT@orcl_11g> select * from test_tab where contains (test_col, 'catalogue') > 0
  2  /
        ID TEST_COL
         1 mail catalog
         2 postal catalogue
         5 catalog
SCOTT@orcl_11g> select * from test_tab where contains (test_col, 'jail') > 0
  2  /
no rows selected
SCOTT@orcl_11g> select * from test_tab where contains (test_col, 'gaol') > 0
  2  /
        ID TEST_COL
         6 jail
         7 gaol
SCOTT@orcl_11g>