| How to obtain key words from the text? [message #76084] |
Tue, 24 February 2004 04:04  |
Valentine
Messages: 1
Registered: February 2004
|
Junior Member |
|
|
|
Can anybody say me how to extract keywords from the document using OracleText. I need to build search engine and use existing documents to formulate new query. There are FREETEXT predicate in the MS SQL. I am looking something like it in the Oracle.
|
|
|
|
| Re: How to obtain key words from the text? [message #76099 is a reply to message #76084] |
Thu, 17 June 2004 09:45  |
Frank Naude
Messages: 4581
Registered: April 1998
|
Senior Member |
|
|
Hi,
Not sure how SQL Server handles it, but in Oracle Text, you can use CTX_DOC.TOKENS() or CTX_QUERY.BROWSE_WORDS() to extract words from an Oracle Text index.
Here's a quick example:
SQL> CREATE TABLE docs (
2 doc_id NUMBER PRIMARY KEY,
3 text CLOB);
Table created.
SQL> INSERT INTO docs VALUES (1, 'Strings to be indexed');
1 row created.
SQL> INSERT INTO docs VALUES (2, 'Second document text');
1 row created.
SQL> COMMIT;
Commit complete.
SQL> CREATE INDEX docs_index ON docs(text) INDEXTYPE IS CTXSYS.CONTEXT;
Index created.
SQL> CREATE TABLE the_tokens (
2 query_id NUMBER,
3 token VARCHAR2(64),
4 offset NUMBER,
5 length NUMBER);
Table created.
SQL> EXEC CTX_DOC.TOKENS('docs_index', '1', 'the_tokens', 1);
PL/SQL procedure successfully completed.
SQL> EXEC CTX_DOC.TOKENS('docs_index', '2', 'the_tokens', 2);
PL/SQL procedure successfully completed.
SQL> COL token FORMAT A30
SQL> SELECT * FROM the_tokens;
QUERY_ID TOKEN OFFSET LENGTH
---------- ------------------------------ ---------- ----------
1 STRINGS 1 7
1 INDEXED 15 7
2 SECOND 1 6
2 DOCUMENT 8 8
2 TEXT 17 4
Best regards.
Frank
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
