DRG-51030: wildcard query expansion after using @

Question

The following query returns one row:
select * from userinterface where contains (searchtx, 'Mustermann within name and %gmx.de within email')>0;
returns one row.
If I add @ (=> [e-mail] within email) the following error occurs:
select * from userinterface where contains (searchtx, 'Mustermann within name and [e-mail] within email')>0;
returns
ERROR at line 1:
ORA-29902: error in executing ODCIIndexStart() routine
ORA-20000: Oracle Text error:
DRG-51030: wildcard query expansion resulted in too many terms
Is the character @ a special character causing the wildcard query expansion? "@gmx.de" is more restictive compared to "gmx.de" (Oracle 10.2.0.4)
select * from userinterface where contains (searchtx, 'Mustermann within name and %de within email')>0;
is also working and returns three rows.

Answer 1

Like Roger said, you could include the @ in your printjoins attribute of your basic_lexer and use a substring_index, then that might allow you to search for '[e-mail]' and find '[email protected]' without encountering the wildcard expansion error. Increasing the wildcard_maxterms would also make the error less likely. However, I believe that is using Oracle Text in a manner other than it is intended and will cause more problems than it will solve, like increasing the size of your domain index tables by storing substrings and making your searches slower.
If you just leave things the way they are, then @ and . are treated as spaces and '[email protected]' is seen as 'somebody gmx de' and indexed as three separate words, so searching for '@gmx.de' will search for 'gmx de' and will find it, so there is no need for a leading wildcard. If you want to avoid problems with errors due to users entering unnecessary leading wildcards with such special characters, then you can replace them in the string before searching. I usually find it convenient to create a cleanup function and include any such problems, then include that in my code that uses a bind variable for the search string. Please see the example below.
SCOTT@orcl_11gR2> -- table:
SCOTT@orcl_11gR2> create table userinterface
  2    (id       number,
  3       searchtx  xmltype)
  4  /
Table created.
SCOTT@orcl_11gR2> -- lexer:
SCOTT@orcl_11gR2> begin
  2    ctx_ddl.create_preference('german_lexer','basic_lexer');
  3    ctx_ddl.set_attribute('german_lexer','composite','german');
  4    ctx_ddl.set_attribute('german_lexer','mixed_case','yes');
  5    ctx_ddl.set_attribute('german_lexer','alternate_spelling','german');
  6  end;
  7  /
PL/SQL procedure successfully completed.
SCOTT@orcl_11gR2> -- index:
SCOTT@orcl_11gR2> CREATE INDEX ui_t_ind on userinterface (searchtx)
  2       indextype is ctxsys.context
  3            PARAMETERS ('
  4                 SECTION GROUP ctxsys.auto_section_group
  5                 LEXER        german_lexer
  6                 MEMORY        100000000
  7                 SYNC        (MANUAL)'
  8            )
  9  /
Index created.
SCOTT@orcl_11gR2> -- insert data:
SCOTT@orcl_11gR2> insert /*+ APPEND */ into userinterface (id, searchtx)
  2  values (1, xmltype (
  3  '<?xml version="1.0"?>
  4  <data>
  5    <name>Mustermann</name>
  6    <email>[email protected]</email>
  7  </data>'))
  8  /
1 row created.
SCOTT@orcl_11gR2> insert /*+ APPEND */ into userinterface (id, searchtx)
  2  values (2, xmltype (
  3  '<?xml version="1.0"?>
  4  <data>
  5    <name>Hans Haeberle</name>
  6    <email>[email protected]</email>
  7  </data>'))
  8  /
1 row created.
SCOTT@orcl_11gR2> -- synchronize, then optimize with rebiuld:
SCOTT@orcl_11gR2> begin
  2    ctx_ddl.sync_index ('ui_t_ind');
  3    ctx_ddl.optimize_index ('ui_t_ind', 'rebuild');
  4  end;
  5  /
PL/SQL procedure successfully completed.
SCOTT@orcl_11gR2> -- what is tokenized, indexed, and searchable:
SCOTT@orcl_11gR2> select token_text from dr$ui_t_ind$i
  2  /
TOKEN_TEXT
DATA
EMAIL
Haeberle
Hans
Häberle
Mann
Muster
Mustermann
NAME
com
de
gmx
somebody
unknown
whatever
15 rows selected.
SCOTT@orcl_11gR2> -- function to clean up search strings:
SCOTT@orcl_11gR2> create or replace function cleanup
  2    (p_string in varchar2)
  3    return         varchar2
  4  as
  5    v_string     varchar2 (100);
  6  begin
  7    v_string := p_string;
  8    v_string := replace (p_string, '%@', ' ');
  9    return v_string;
10  end cleanup;
11  /
Function created.
SCOTT@orcl_11gR2> show errors
No errors.
SCOTT@orcl_11gR2> -- example search strings, queries, and results:
SCOTT@orcl_11gR2> column name  format a15
SCOTT@orcl_11gR2> column email format a20
SCOTT@orcl_11gR2> variable search_string varchar2 (100)
SCOTT@orcl_11gR2> begin
  2    :search_string :=
  3        'Mustermann within name and [e-mail] within email';
  4  end;
  5  /
PL/SQL procedure successfully completed.
SCOTT@orcl_11gR2> select id,
  2           extractvalue (u.searchtx, '//name') name,
  3           extractvalue (u.searchtx, '//email') email
  4  from   userinterface u
  5  where  contains (searchtx, cleanup (:search_string)) > 0
  6  /
        ID NAME            EMAIL
         1 Mustermann      [email protected]
1 row selected.
SCOTT@orcl_11gR2> begin
  2    :search_string :=
  3        'Häberle within name and unknown@whatever within email';
  4  end;
  5  /
PL/SQL procedure successfully completed.
SCOTT@orcl_11gR2> select id,
  2           extractvalue (u.searchtx, '//name') name,
  3           extractvalue (u.searchtx, '//email') email
  4  from   userinterface u
  5  where  contains (searchtx, cleanup (:search_string)) > 0
  6  /
        ID NAME            EMAIL
         2 Hans Haeberle   [email protected]
1 row selected.
SCOTT@orcl_11gR2>