full text index 探秘（1）-优快云博客

本文通过具体实验展示了Oracle数据库中全文索引的工作原理及性能影响，特别是在高频率DML操作场景下的表现，并探讨了不同优化策略的效果。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

[oracle@roger ~]$ sqlplus "/as sysdba"

SQL*Plus: Release 10.2.0.4.0 - Production on Sun Aug 28 17:17:45 2011

Copyright (c) 1982, 2007, Oracle. All Rights Reserved.

Connected to an idle instance.

SQL> startup
ORACLE instance started.

Total System Global Area 167772160 bytes
Fixed Size 1266392 bytes
Variable Size 71306536 bytes
Database Buffers 92274688 bytes
Redo Buffers 2924544 bytes
Database mounted.
Database opened.
SQL> create user text_idx identified by text_idx;

User created.

SQL>
SQL> grant resource ,connect,ctxapp to text_idx;

Grant succeeded.

SQL>
SQL> GRANT EXECUTE ON CTXSYS.CTX_CLS TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_DDL TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_DOC TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_OUTPUT TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_QUERY TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_REPORT TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_THES TO text_idx;
GRANT EXECUTE ON CTXSYS.CTX_ULEXER TO text_idx;

Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
Grant succeeded.

SQL>
SQL> conn text_idx/text_idx
Connected.
SQL> CREATE TABLE docs (id NUMBER PRIMARY KEY, text VARCHAR2(500));

Table created.
SQL> INSERT INTO docs VALUES(1, 'http://www.itpub.net/thread-1475450-1-1.html');

1 row created.

SQL> INSERT INTO docs VALUES(2, 'http://www.yesky.com/285/1942785_1.shtml');
INSERT INTO docs VALUES(3, 'http://www.baidu.com/s?tn=chenly082=4391');

1 row created.

SQL> INSERT INTO docs VALUES(4, 'file:///D:/oracle---官方文档/B19306_01/B19306_01/text.102/b14217/quicktour.htmi1008362');
1 row created.

SQL>
INSERT INTO docs VALUES(5, 'file:///D:/oracle---官方文档/B19306_01/B19306_01/text.10
1 row created.

SQL> 2/b14217/acase.htm#g637598');
INSERT INTO docs VALUES(6, 'http://database.51cto.com/art/201104/252898.htm');

1 row created.

SQL>
1 row created.

SQL>
SQL> commit;

Commit complete.

SQL> select count(*) from docs;

 COUNT(*)
----------
 6

SQL> col text for a50
SQL> set lines 120
SQL> SELECT id, text FROM docs WHERE CONTAINS(text, 'baidu', 1) > 0;

 ID TEXT
---------- --------------------------------------------------
 3 http://www.baidu.com/s?tn=chenly082=4391

SQL>
SQL> set autot traceonly
SQL> SELECT id, text FROM docs WHERE CONTAINS(text, 'baidu', 1) > 0;

Execution Plan
----------------------------------------------------------
Plan hash value: 3588628665

----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 277 | 4 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| DOCS | 1 | 277 | 4 (0)| 00:00:01 |
|* 2 | DOMAIN INDEX | IDX_DOCS | | | 4 (0)| 00:00:01 |
----------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

 2 - access("CTXSYS"."CONTAINS"("TEXT",'baidu',1)>0)

Note
-----
 - dynamic sampling used for this statement

Statistics
----------------------------------------------------------
 11 recursive calls
 0 db block gets
 18 consistent gets
 0 physical reads
 0 redo size
 500 bytes sent via SQL*Net to client
 400 bytes received via SQL*Net from client
 2 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 1 rows processed

SQL> create sequence seq_docs
 2 minvalue 1
 3 nomaxvalue
 4 start with 7
 5 increment by 1
 6 nocycle
 7 cache 10;

Sequence created.

SQL> begin
 2 for i in 1..100 loop
 3 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/');
 4 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=210');
 5 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=216');
 6 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=201');
 7 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=196');
 8 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=219');
 9 commit;
10 end loop;
11 end;
12 /

PL/SQL procedure successfully completed.

SQL>
SQL>
SQL> select count(*) from docs;

 COUNT(*)
----------
 606

SQL> begin
 2 for i in 1..1000 loop
 3 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/');
 4 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=210');
 5 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=216');
 6 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=201');
 7 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=196');
 8 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=219');
 9 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=205');
10 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=206');
11 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=207');
12 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=199');
13 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=208');
14 commit;
15 end loop;
16 end;
17 /

PL/SQL procedure successfully completed.

SQL>
SQL> select count(*) from docs;

 COUNT(*)
----------
 11606

SQL>
SQL> set timing on
SQL> begin
 2 for i in 1..10000 loop
 3 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/');
 4 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=210');
 5 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=216');
 6 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=201');
 7 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=196');
 8 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=219');
 9 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=205');
10 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=206');
11 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=207');
12 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=199');
13 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=208');
14 commit;
15 end loop;
16 end;
17 /

PL/SQL procedure successfully completed.

Elapsed: 00:03:43.23
SQL>
SQL> select count(*) from docs;

 COUNT(*)
----------
 121606

Elapsed: 00:00:00.05
SQL>
SQL> select count(*) from docs where CONTAINS(text,'p=210') > 0;

 COUNT(*)
----------
 0

SQL> select count(*) from docs where CONTAINS(text,'killdb') >0;

 COUNT(*)
----------
 0

SQL> EXEC CTX_DDL.SYNC_INDEX('idx_docs', '5m');

PL/SQL procedure successfully completed.

SQL> select count(*) from docs where CONTAINS(text,'p=210') > 0;

 COUNT(*)
----------
 110500

SQL>
SQL> SELECT count(*) FROM docs WHERE CONTAINS(text,'killdb') > 0;

 COUNT(*)
----------
 121600

从这里看出 同步index的重要性了,下面来创建job 来进行定期的同步index以及优化。
SQL> create or replace procedure sync_idx_docs as
 2 begin
 3 ctx_ddl.sync_index('idx_docs');
 4 end;
 5 /

Procedure created.

SQL> VARIABLE job_no number;
SQL> BEGIN
 2 DBMS_JOB.SUBMIT(:job_no,'sync_idx_docs();',
 3 SYSDATE, 'SYSDATE + 1');
 4 commit;
 5 END;
 6 /

PL/SQL procedure successfully completed.

SQL> create or replace procedure optimize_idx_docs as
 2 begin
 3 ctx_ddl.optimize_index('idx_docs','FULL');
 4 end;
 5 /

Procedure created.

SQL> VARIABLE job_no number;
SQL> BEGIN
 2 DBMS_JOB.SUBMIT(:job_no,'optimize_idx_docs();',
 3 SYSDATE, 'SYSDATE + 1');
 4 commit;
 5 END;
 6 /

PL/SQL procedure successfully completed.

SQL>
SQL> col INTERVAL for a25
SQL> select JOB,LOG_USER,SCHEMA_USER,INTERVAL,LAST_DATE from user_jobs;

 JOB LOG_USER SCHEMA_USER INTERVAL LAST_DATE
---------- ------------- --------------- ---------------- ---------
 21 TEXT_IDX TEXT_IDX SYSDATE + 1 28-AUG-11
 22 TEXT_IDX TEXT_IDX SYSDATE + 1 28-AUG-11

---启动如上2个job任务
SQL> execute dbms_job.run(21);
execute dbms_job.run(22);
PL/SQL procedure successfully completed.

SQL>

PL/SQL procedure successfully completed.

SQL>
SQL> alter table docs add owner varchar2(10);

Table altered.

SQL> update docs set owner='Roger';

121606 rows updated.

SQL> commit;

Commit complete.

SQL>

下面来测试，如果表docs dml操作比较频繁，那么对于select语句来说，是否有较大的影响？

SQL> select /*+ no_index(docs SYS_C005195)*/
 2 count(*)
 3 from docs
 4 where contains(text, 'p=199') > 0
 5 and id > 10500
 6 and id < 10800;
Elapsed: 00:00:00.09

Execution Plan
----------------------------------------------------------
Plan hash value: 1092983528

-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 277 | 4 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 277 | | |
|* 2 | TABLE ACCESS BY INDEX ROWID| DOCS | 1 | 277 | 4 (0)| 00:00:01 |
|* 3 | DOMAIN INDEX | IDX_DOCS | | | 4 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

 2 - filter("ID">10500 AND "ID"<10800)
 3 - access("CTXSYS"."CONTAINS"("TEXT",'p=199')>0)

Note
-----
 - dynamic sampling used for this statement

---session 1
SQL> show user
USER is "TEXT_IDX"
SQL> begin
 2 for i in 1 .. 15000 loop
 3 if mod(i, 2) = 0 then
 4 update docs
 5 set owner = 'killdb'
 6 where id = i
 7 and id > 10000
 8 and id < 15000;
 9 commit;
10 end if;
11 end loop;
12 commit;
13 end;
14 /

PL/SQL procedure successfully completed.

---session 2
SQL> select /*+ no_index(docs SYS_C005195)*/
 2 id,owner,text
 3 from docs
 4 where contains(text, 'p=199') > 0
 5 and id > 10700
 6 and id < 10800;

90 rows selected.

Elapsed: 00:00:00.19

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 17447 consistent gets
 0 physical reads
 1024 redo size
 4691 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.35

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 21322 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

/90 rows selected.

Elapsed: 00:00:00.76

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 19518 consistent gets
 0 physical reads
 1856 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL>

90 rows selected.

Elapsed: 00:00:00.37

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15409 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.30

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 20226 consistent gets
 0 physical reads
 64 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.32

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 17881 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.42

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15728 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.49

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 17406 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.25

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 16768 consistent gets
 0 physical reads
 128 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.40

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 16515 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.34

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 16046 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.34

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 19300 consistent gets
 0 physical reads
 1856 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:02.92

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 19860 consistent gets
 0 physical reads
 1216 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.45

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15569 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.42

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15120 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.21

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 17201 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.14

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 18365 consistent gets
 0 physical reads
 384 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.59

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 17998 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.29

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 14515 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.11

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15595 consistent gets
 0 physical reads
 704 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.23

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15675 consistent gets
 0 physical reads
 768 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.20

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 15584 consistent gets
 0 physical reads
 0 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL> /

90 rows selected.

Elapsed: 00:00:00.14

Statistics
----------------------------------------------------------
 13 recursive calls
 0 db block gets
 18380 consistent gets
 0 physical reads
 1344 redo size
 4689 bytes sent via SQL*Net to client
 455 bytes received via SQL*Net from client
 7 SQL*Net roundtrips to/from client
 0 sorts (memory)
 0 sorts (disk)
 90 rows processed

SQL>

从上面的测试不难看出，对于DML操作频繁的话，对全文索引来说，还是有较大的影响的。

SQL> begin
 2 for i in 1 .. 15000 loop
 3 if mod(i, 2) = 0 then
 4 update docs
 5 set owner = 'google'
 6 where id = i
 7 and id > 50000
 8 and id < 10000;
 9 commit;
10 end if;
11 end loop;
12 commit;
13 end;
14 /

PL/SQL procedure successfully completed.

Elapsed: 00:00:02.25
SQL> exec ctx_ddl.optimize_index('idx_docs','rebuild') ;

PL/SQL procedure successfully completed.

Elapsed: 00:00:02.84
SQL> begin
 2 for i in 1 .. 15000 loop
 3 if mod(i, 2) = 0 then
 4 update docs
 5 set owner = 'baidu'
 6 where id = i
 7 and id > 50000
 8 and id < 10000;
 9 commit;
10 end if;
11 end loop;
12 commit;
13 end;
14 /

PL/SQL procedure successfully completed.

Elapsed: 00:00:02.23
SQL> exec ctx_ddl.optimize_index('idx_docs','FULL') ;

PL/SQL procedure successfully completed.

Elapsed: 00:00:00.45
SQL>

SQL> begin
 2 for i in 1..10000 loop
 3 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/');
 4 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=210');
 5 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=216');
 6 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=201');
 7 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=196');
 8 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=219');
 9 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=205');
10 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=206');
11 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=207');
12 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=199');
13 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=208');
14 commit;
15 end loop;
16 end;
17 /

begin
*
ERROR at line 1:
ORA-29875: failed in the execution of the ODCIINDEXINSERT routine
ORA-20000: Oracle Text error:
DRG-50857: oracle error in textindexmethods.ODCIIndexInsert
ORA-00604: error occurred at recursive SQL level 2
ORA-01013: user requested cancel of current operation
ORA-06512: at "CTXSYS.DRUE", line 160
ORA-06512: at "CTXSYS.TEXTINDEXMETHODS", line 633
ORA-06512: at line 13

Elapsed: 00:01:09.73

SQL> select count(*) from docs;

 COUNT(*)
----------
 138898

Elapsed: 00:00:00.35
SQL> exec ctx_ddl.optimize_index('idx_docs','FULL') ;

PL/SQL procedure successfully completed.

Elapsed: 00:00:00.48
SQL> begin
 2 for i in 1..10000 loop
 3 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/');
 4 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=210');
 5 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=216');
 6 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=201');
 7 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=196');
 8 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=219');
 9 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=205');
10 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=206');
11 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=207');
12 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=199');
13 INSERT INTO docs(id, text) VALUES(seq_docs.NEXTVAL, 'http://www.killdb.com/?p=208');
14 commit;
15 end loop;
16 end;
17 /

begin
*
ERROR at line 1:
ORA-29875: failed in the execution of the ODCIINDEXINSERT routine
ORA-20000: Oracle Text error:
DRG-50857: oracle error in textindexmethods.ODCIIndexInsert
ORA-01013: user requested cancel of current operation
ORA-06512: at "CTXSYS.DRUE", line 160
ORA-06512: at "CTXSYS.TEXTINDEXMETHODS", line 633
ORA-06512: at line 12

Elapsed: 00:01:06.17

SQL> exec ctx_ddl.optimize_index('idx_docs','rebuild') ;

PL/SQL procedure successfully completed.

Elapsed: 00:00:02.09
SQL> select count(*) from docs;

 COUNT(*)
----------
 156091

Elapsed: 00:00:00.01
SQL>
有人说优化index的时候，使用rebuild比full快很多，但是我测试恰恰相反，不知道为啥。

最后总结下：

1. DML操作较为频繁表，对全文索引一定影响，但是如果说必须使用该功能，那么也没办法,
2. 至于说为什么有一定影响，其实原理都一样的，因为逻辑读的消耗会增加，必然影响性能。

该功能是从oracle 9i引入的，10g，11g都分别有一定的改变，下一篇文章将会进行描述。