Help needed with oracle text special character search

Question

Hi all
Using oracle 11g sql developer 4.0
I am facing this challenge where Oracle text when it comes to searching text that contains special character.
This what I have done so far with help of http://www.orafaq.com/forum/t/162229/
  CREATE TABLE "SOS"."COMPANY"
   ( "COMPANY_ID" NUMBER(10,0) NOT NULL ENABLE,
  "COMPANY_NAME" VARCHAR2(50 BYTE),
  "ADDRESS1" VARCHAR2(50 BYTE),
  "ADDRESS2" VARCHAR2(10 BYTE),
  "CITY" VARCHAR2(40 BYTE),
  "STATE" VARCHAR2(20 BYTE),
  "ZIP" NUMBER(5,0)
   ) SEGMENT CREATION IMMEDIATE
  PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 NOCOMPRESS LOGGING
  STORAGE(INITIAL 65536 NEXT 1048576 MINEXTENTS 1 MAXEXTENTS 2147483645
  PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1 BUFFER_POOL DEFAULT FLASH_CACHE DEFAULT CELL_FLASH_CACHE DEFAULT)
  TABLESPACE "USERS" ;
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (1,'LSG SOLUTIONS LLC',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (2,'LOVE''S TRAVEL',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (3,'DEVON ENERGY',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (4,'SONIC INC',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (5,'MSCI',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (6,'ERNEST AND YOUNG',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (7,'JOHN DEER',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (8,'Properties@Oklahoma, LLC',null,null,null,null,null);
Insert into COMPANY (COMPANY_ID,COMPANY_NAME,ADDRESS1,ADDRESS2,CITY,STATE,ZIP) values (9,'D.D.T  L.L.C.',null,null,null,null,null);
   BEGIN
CTX_DDL.CREATE_PREFERENCE ('your_lexer', 'BASIC_LEXER');
     CTX_DDL.SET_ATTRIBUTE ('your_lexer', 'SKIPJOINS', '.,@-'''); -- to skip . , @ - ' symbols
    END;
  CREATE INDEX my_index2 ON COMPANY(COMPANY_NAME)
     INDEXTYPE IS CTXSYS.CONTEXT PARALLEL
   PARAMETERS ('LEXER your_lexer');   
SELECT
company_name
FROM company
WHERE CATSEARCH(company.COMPANY_NAME, 'LLC','') > 0
ORDER BY company.COMPANY_ID;
output
company_name
1 LSG SOLUTIONS LLC
2 Properties@Oklahoma, LLC
only return 2 row but should return 3

Answer 1

I just noticed that I forgot to use an empty stoplist, so I have added that to the revised example below.  Otherwise, it uses a default stoplist that would not index common single-letter words like A and I.
1. Whtat is Just search on single character 'L'? It give me error.
Since it uses the NEAR operator, searching for just one letter causes incomplete syntax, asking it to search for L near a missing second value.  So, I have added additional code to allow for just one letter.
2. How do I do auto refresh on this index on datastore?
If I add "sync        (on commit)" it does not refresh the previously set token.
Sync(on commit) does synchronize so that the data is immediately searchable.  You have to either optimize or rebuild or drop and recreate the index to condense the rows in the domain index table.
3.lastly explanation of
<seq>NEAR((' || letters_func (:search_string) || '),1,TRUE)</seq>
                  <seq>NEAR((' || letters_func (:search_string) || '),100,TRUE)</seq>
                <seq>NEAR((' || letters_func (:search_string) || '),100,FALSE)</seq>
why 100 true and 100 false
100 is just a default value that I used for the second parameter of near, indicating how close the letters need to be to each other.  True and False are values for the third parameter of near, indicating whether or not the letters must be in the same order or not.  So, it returns the results in the order of first those that are very close to one another and in the same order, then those that may be further away but in the same order, then those that may be further away and in any order.
SCOTT@orcl12c> CREATE TABLE company_near
  2    (company_id    NUMBER(10,0) NOT NULL ENABLE,
  3      company_name  VARCHAR2(50 BYTE))
  4  /
Table created.
SCOTT@orcl12c> SET DEFINE OFF
SCOTT@orcl12c> BEGIN
  2  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (1,'LSG SOLUTIONS LLC');
  3  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (2,'LOVE''S TRAVEL');
  4  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (3,'DEVON ENERGY');
  5  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (4,'SONIC INC');
  6  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (5,'MSCI');
  7  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (6,'ERNEST AND YOUNG');
  8  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (7,'JOHN DEER');
  9  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (8,'Properties@Oklahoma, LLC');
10  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (9,'D.D.T  L.L.C.');
11  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (10,'LSG COMPANY, LLC');
12  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (11,'LSG STAFFING, LLC');
13  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (12,'L & S GROUP LLC');
14  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (13,'L S & G, INC.');
15  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (14,'L.S.G. PROPERTIES, L.L.C.');
16  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (15,'LSGS PROPERTIES, LLC');
17  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (16,'LSQ INVESTORS, L.L.C');
18  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (17,'LHP SHERMAN/GRAYSON, LLC');
19  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (18,'Walmart');
20  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (19,'Wal mart');
21  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (20,'LSG Property Investments, L.L.C.');
22  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (21,'1224 S GALVESTON AVE, LLC');
23  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (22,'1527 S GARY AVE, LLC');
24  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (23,'FIFTEENTH STREET GRILL');
25  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (24,'Massa Lobortis LLP');
26  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (25,'Risus A Inc.');
27  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (26,'Dollar $ store');
28  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (27,'L.O.V.E., INC. ');
29  Insert into COMPANY_NEAR (COMPANY_ID,COMPANY_NAME) values (28,'J-MART LLC ');
30  END;
31  /
PL/SQL procedure successfully completed.
SCOTT@orcl12c> CREATE OR REPLACE FUNCTION letters_func
  2    (p_string IN VARCHAR2)
  3    RETURN        VARCHAR2
  4  AS
  5    v_string     VARCHAR2(4000);
  6  BEGIN
  7    FOR i IN 1 .. LENGTH (p_string)
  8    LOOP
  9       IF REGEXP_LIKE (SUBSTR (p_string, i, 1), '[A-Z]', 'i')
10       THEN
11         v_string := v_string || SUBSTR (p_string, i, 1) || ',';
12       END IF;
13    END LOOP;
14    v_string := RTRIM (v_string, ',');
15    RETURN v_string;
16  END letters_func;
17  /
Function created.
SCOTT@orcl12c> BEGIN
  2    CTX_DDL.CREATE_PREFERENCE ('letters_datastore', 'MULTI_COLUMN_DATASTORE');
  3    CTX_DDL.SET_ATTRIBUTE
  4       ('letters_datastore',
  5        'COLUMNS',
  6        'letters_func (company_name) company_name');
  7    CTX_DDL.SET_ATTRIBUTE ('letters_datastore', 'DELIMITER', 'NEWLINE');
  8  END;
  9  /
PL/SQL procedure successfully completed.
SCOTT@orcl12c> CREATE INDEX letters_index ON company_near (company_name)
  2  INDEXTYPE IS CTXSYS.CONTEXT
  3  PARAMETERS
  4    ('DATASTORE  letters_datastore
  5       STOPLIST   CTXSYS.EMPTY_STOPLIST
  6       SYNC        (ON COMMIT)')
  7  /
Index created.
SCOTT@orcl12c> SELECT COUNT(*) FROM dr$letters_index$i
  2  /
  COUNT(*)
        24
1 row selected.
SCOTT@orcl12c> VARIABLE search_string VARCHAR2(100)
SCOTT@orcl12c> EXEC :search_string := 'LSG'
PL/SQL procedure successfully completed.
SCOTT@orcl12c> SELECT SCORE(1), company_id, company_name
  2  FROM   company_near
  3  WHERE  CONTAINS
  4            (company_name,
  5             '<query>
  6            <textquery>
  7              <progression>
  8                <seq>'       || :search_string            || '</seq>
  9                <seq>NEAR((' || letters_func (:search_string) || '),1,TRUE)</seq>
10                <seq>NEAR((' || letters_func (:search_string) || '),100,TRUE)</seq>
11                <seq>NEAR((' || letters_func (:search_string) || '),100,FALSE)</seq>
12              </progression>
13            </textquery>
14             </query>',
15             1) > 0
16  ORDER  BY SCORE(1) DESC
17  /
  SCORE(1) COMPANY_ID COMPANY_NAME
        56          1 LSG SOLUTIONS LLC
        56         10 LSG COMPANY, LLC
        56         11 LSG STAFFING, LLC
        56         12 L & S GROUP LLC
        56         13 L S & G, INC.
        56         14 L.S.G. PROPERTIES, L.L.C.
        56         20 LSG Property Investments, L.L.C.
        56         15 LSGS PROPERTIES, LLC
        31         17 LHP SHERMAN/GRAYSON, LLC
         8         21 1224 S GALVESTON AVE, LLC
         4         22 1527 S GARY AVE, LLC
         4         23 FIFTEENTH STREET GRILL
12 rows selected.
SCOTT@orcl12c> EXEC :search_string := 'L'
PL/SQL procedure successfully completed.
SCOTT@orcl12c> /
  SCORE(1) COMPANY_ID COMPANY_NAME
        78          1 LSG SOLUTIONS LLC
        77          8 Properties@Oklahoma, LLC
        77          9 D.D.T  L.L.C.
        77         10 LSG COMPANY, LLC
        77         11 LSG STAFFING, LLC
        77         12 L & S GROUP LLC
        77         28 J-MART LLC
        77          2 LOVE'S TRAVEL
        77         26 Dollar $ store
        77         24 Massa Lobortis LLP
        77         23 FIFTEENTH STREET GRILL
        77         14 L.S.G. PROPERTIES, L.L.C.
        77         15 LSGS PROPERTIES, LLC
        77         16 LSQ INVESTORS, L.L.C
        77         17 LHP SHERMAN/GRAYSON, LLC
        77         20 LSG Property Investments, L.L.C.
        77         21 1224 S GALVESTON AVE, LLC
        77         22 1527 S GARY AVE, LLC
        76         19 Wal mart
        76         18 Walmart
        76         27 L.O.V.E., INC.
        76         13 L S & G, INC.
22 rows selected.
SCOTT@orcl12c> INSERT INTO company_near (company_id, company_name) VALUES (30, 'Laris Gordman llc.'  )
  2  /
1 row created.
SCOTT@orcl12c> COMMIT
  2  /
Commit complete.
SCOTT@orcl12c> SELECT COUNT(*) FROM dr$letters_index$i
  2  /
  COUNT(*)
        35
1 row selected.
SCOTT@orcl12c> EXEC :search_string := 'Laris Gordman llc.'
PL/SQL procedure successfully completed.
SCOTT@orcl12c> SELECT SCORE(1), company_id, company_name
  2  FROM   company_near
  3  WHERE  CONTAINS
  4            (company_name,
  5             '<query>
  6            <textquery>
  7              <progression>
  8                <seq>NEAR((' || letters_func (:search_string) || '),1,TRUE)</seq>
  9                <seq>NEAR((' || letters_func (:search_string) || '),100,TRUE)</seq>
10                <seq>NEAR((' || letters_func (:search_string) || '),100,FALSE)</seq>
11              </progression>
12            </textquery>
13             </query>',
14             1) > 0
15  ORDER  BY SCORE(1) DESC
16  /
  SCORE(1) COMPANY_ID COMPANY_NAME
       100         30 Laris Gordman llc.
1 row selected.
SCOTT@orcl12c> EXEC CTX_DDL.OPTIMIZE_INDEX ('letters_index', 'FULL')
PL/SQL procedure successfully completed.
SCOTT@orcl12c> SELECT COUNT(*) FROM dr$letters_index$i
  2  /
  COUNT(*)
        24
1 row selected.