@
目錄
一、Oracle索引簡介
在看《收獲，不止SQL優化》一書，並根據書中例子進行實踐，整理成筆記
1.1 索引分類
Oracle索引分為BTree索引、位圖索引、反向索引、函數索引、全文索引等等。
1.2 索引數據結構
Oracle索引中最常用的是BTree索引，所以就以BTree索引為例，講一下BTree索引，BTree索引數據結構是一種二叉樹的結構，索引由根塊(Root)、莖塊(Branch)、葉子塊(Leaf)組成，其中葉子塊主要存儲索引列具體值(Key Column Value)以及能定位到數據塊具體位置的Rowid，莖塊和根塊主要保存對應下級對應索引
1.3 索引特性
索引特性：
- 索引本身是有序的
- 索引本身能存儲列值
1.4 索引使用注意要點
- (1)、僅等值無范圍查詢時，組合的順序不影晌性能
drop table t purge;
create table t as select * from dba objects;
update t set object_id=rownum ;
commit;
create index idx_id_type on t(object_id, object_type) ;
create index idx_type_id on t(object_type , object_id) ;
set autotrace off;
alter session set statistics_level=all ;
select /*+index(t idx_id_type)*/ * from t where object_id=20 and object_type='TABLE';
select * from table(dbms_xplan.display cursor(null , null , 'allstats last'));
select /*+index(t,idx_type id)*/ * from t where object_id=20 and object_type= 'TABLE';
select * from table(dbms_xplan.display cursor(null , null , 'allstats last'));
- (2)、范圍查詢時，組合索引最佳順序一般是將等值查詢的列置前
select /*+index (t, idx_id_type)*/ * from t where object_id>=20 and object_id<2000 and
object_type='TABLE';
select /*+index (t , idx_type_id) */ * from t where object_id>=20 and object_id<2000
and object type='TABLE';
- (3)、Oracle不能同時在索引根的兩段尋找最大值和最小值
set autotrace on
select max(object_id) , min(object_id) from t;
笛卡爾乘積寫法：
set autotrace on
select max, min
from (select max(object_id) max from t ) a ,
(select min(object_id) min from t ) b;
- (4)、索引最新的數據塊一般是在最右邊
1.5、索引的缺點
- 熱快競爭：索引最新的數據塊一般在最右邊，而訪問也一般是訪問比較新的數據，所以容易造成熱快競爭
- 更新新增問題：索引本身是有序的，所以查詢時候很快，但是更新時候就麻煩了，新增更新索引都需要保證排序
1.6、索引失效
索引失效分為邏輯失效和物理失效
- 邏輯失效
邏輯失效是因為一些sql語法導致索引失效，比如加了一些函數，而索引列不是函數索引 - 物理失效
物理失效是真的失效,比如被設置unusable屬性,分區表的不規范操作也會導致索引失效等等情況
alter index index_name unusable;
二、索引分類介紹
索引分類：BTree索引、位圖索引、函數索引、反向索引、全文索引
2.1、位圖索引
位圖索引：位圖索引儲存的就是比特值
環境准備，位圖索引性質適用於count時，效率最高
drop table t purge;
create table t as select * from dba_objects;
update t set object_id = rownum;
commit;
不用索引的情況：
SQL> set autotrace on
SQL> select count(*) from t;
COUNT(*)
----------
72016
執行計划
----------------------------------------------------------
Plan hash value: 2966233522
-------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 288 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| T | 86565 | 288 (1)| 00:00:04 |
-------------------------------------------------------------------
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
4 recursive calls
0 db block gets
1111 consistent gets
0 physical reads
0 redo size
432 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
創建位圖索引：
create bitmap index idx_bitm_t_status on t(status);
再次查詢，走位圖索引查詢：
SQL> set autotrace on
SQL> select count(*) from t;
COUNT(*)
----------
72016
執行計划
----------------------------------------------------------
Plan hash value: 4272013625
--------------------------------------------------------------------------------
-----------
| Id | Operation | Name | Rows | Cost (%CPU)|
Time |
--------------------------------------------------------------------------------
-----------
| 0 | SELECT STATEMENT | | 1 | 5 (0)|
00:00:01 |
| 1 | SORT AGGREGATE | | 1 | |
|
| 2 | BITMAP CONVERSION COUNT | | 86565 | 5 (0)|
00:00:01 |
| 3 | BITMAP INDEX FAST FULL SCAN| IDX_BITM_T_STATUS | | |
|
--------------------------------------------------------------------------------
-----------
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
0 recursive calls
0 db block gets
6 consistent gets
0 physical reads
0 redo size
432 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
注意要點：
位圖索引更新列容易造成死鎖，所以查詢比較多列才適合建位圖索引，更新比較多的列就盡量不要建索引
1.2、函數索引
函數索引：就是將一個函數計算的結果存儲在行的列中
環境准備：
drop table t purge;
create table t (id int, status varchar2(2));
insert into t select 1,'N' from dual;
insert into t select rownum ,'Y' from dual connect by rownum <1000;
commit;
不走索引的查詢：
SQL> set autotrace on
SQL> select * from t where (case when status='N' then 'No' end)='No';
ID STAT
---------- ----
1 N
執行計划
----------------------------------------------------------
Plan hash value: 1601196873
--------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 16 | 3 (0)| 00:00:01 |
|* 1 | TABLE ACCESS FULL| T | 1 | 16 | 3 (0)| 00:00:01 |
--------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter(CASE "STATUS" WHEN 'N' THEN 'No' END ='No')
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
0 recursive calls
0 db block gets
8 consistent gets
0 physical reads
0 redo size
486 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
創建函數索引：
create index idx_status on t (case when status ='N' then 'No' end);
走函數索引的查詢：
SQL> select * from t where (case when status='N' then 'No' end)='No';
ID STAT
---------- ----
1 N
執行計划
----------------------------------------------------------
Plan hash value: 3908194542
--------------------------------------------------------------------------------
----------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)|
Time |
--------------------------------------------------------------------------------
----------
| 0 | SELECT STATEMENT | | 10 | 200 | 2 (0)|
00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T | 10 | 200 | 2 (0)|
00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_STATUS | 4 | | 1 (0)|
00:00:01 |
--------------------------------------------------------------------------------
----------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access(CASE "STATUS" WHEN 'N' THEN 'No' END ='No')
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
0 recursive calls
0 db block gets
2 consistent gets
0 physical reads
0 redo size
486 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL>
注意要點：
自定義函數時要加上deterministic 關鍵字，不然不能建立函數索引
建立一個自定義函數：
create or replace function f_addusl(i int) return int is
begin
return(i + 1);
end;
嘗試建立函數索引：
create index idx_ljb_test on t(f_addusl(id));
提示：ORA-30553:函數不能確定
用deterministic 關鍵字，就可以建立函數索引
create or replace function f_addusl(i int) return int deterministic is
begin
return(i + 1);
end;
在自定義函數代碼更新時，對應的函數索引也要重建，否則不能用到原來的函數索引
1.3、反向索引
反向索引：反向索引其實也是BTree索引的一種特例，不過在列中字節會反轉的(反向索引是為了避免熱快競爭，比如索引列中存儲的列值是遞增的，比如250101,250102，按照BTree索引的特性，一般是按照順序存儲在索引右邊的，所以容易形成熱快競爭，而反向索引可以避免這種情況，因為反向索引是這樣存儲的，比如101052,201052，這樣列值就距離很遠了，避免了熱快競爭)
反向索引不能用到范圍查詢
SQL> set autotrace on
SQL> select * from t where created=sysdate;
未選定行
執行計划
----------------------------------------------------------
Plan hash value: 913247507
--------------------------------------------------------------------------------
---------------
| Id | Operation | Name | Rows | Bytes | Cost (%C
PU)| Time |
--------------------------------------------------------------------------------
---------------
| 0 | SELECT STATEMENT | | 12 | 2484 | 286
(0)| 00:00:04 |
| 1 | TABLE ACCESS BY INDEX ROWID| T | 12 | 2484 | 286
(0)| 00:00:04 |
|* 2 | INDEX RANGE SCAN | IDX_REV_CREATED | 346 | | 1
(0)| 00:00:01 |
--------------------------------------------------------------------------------
---------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("CREATED"=SYSDATE@!)
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
0 recursive calls
0 db block gets
2 consistent gets
0 physical reads
0 redo size
1191 bytes sent via SQL*Net to client
408 bytes received via SQL*Net from client
1 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
0 rows processed
SQL>
范圍查詢，發現不走反向索引查詢
SQL> select * from t where created>= sysdate-10 and created <= sysdate-1;
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
SYS
ICOL$
20
2 TABLE 15-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
1
SYS
I_USER1
46
46 INDEX 14-6月 -19
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
02-4月 -10 2010-04-02:13:18:38 VALID N N N
4
SYS
CON$
28
28 TABLE 13-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
1
SYS
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
UNDO$
15
15 TABLE 12-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
1
SYS
C_COBJ#
29
29 CLUSTER 11-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
5
SYS
I_OBJ#
3
3 INDEX 10-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
4
SYS
PROXY_ROLE_DATA$
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
25
25 TABLE 09-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
1
SYS
I_IND1
41
41 INDEX 08-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
4
OWNER
------------------------------------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
SUBOBJECT_NAME OBJECT_ID
------------------------------------------------------------ ----------
DATA_OBJECT_ID OBJECT_TYPE CREATED
-------------- -------------------------------------- --------------
LAST_DDL_TIME TIMESTAMP STATUS TE GE SE
-------------- -------------------------------------- -------------- -- -- --
NAMESPACE EDITION_NAME
---------- ------------------------------------------------------------
SYS
I_CDEF2
54
54 INDEX 07-6月 -19
02-4月 -10 2010-04-02:13:18:38 VALID N N N
4
已選擇9行。
執行計划
----------------------------------------------------------
Plan hash value: 1322348184
---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 12 | 2484 | 292 (2)| 00:00:04 |
|* 1 | FILTER | | | | | |
|* 2 | TABLE ACCESS FULL| T | 12 | 2484 | 292 (2)| 00:00:04 |
---------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter(SYSDATE@!-10<=SYSDATE@!-1)
2 - filter("CREATED">=SYSDATE@!-10 AND "CREATED"<=SYSDATE@!-1)
Note
-----
- dynamic sampling used for this statement (level=2)
統計信息
----------------------------------------------------------
5 recursive calls
0 db block gets
1112 consistent gets
0 physical reads
0 redo size
1770 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
9 rows processed
SQL>
1.4、全文索引
全文索引：所謂Oracle全文索引是通過Oracle詞法分析器(lexer)將所有的表意單元term存儲dr$開頭的表里並存儲term出現的位置、次數、hash值等等信息，Oracle提供了basic_lexer(針對英語)、chinese_vgram_lexer(漢語分析器)、chinese_lexer(新的漢語分析器)。
- basic_lexer：是一種適用於英文的分析器，根據空格或者標點符號將詞元分離，不管對於中文來說是沒有空格的，所以這種分析器不適合中文
- chinese_vgram_lexer：這是一種原先專門的中文分析器，支持所有的漢字字符集，比如zhs16gbk單點。這種分析器，分析過程是按字為單元進行分析的，舉個例子，“索引本身是有序的”，按照這種分析器，會分成詞元“索”、“索引”、“引本”、“本身”、“身是”、“是有”、“有序”、“序的”、“的”這些詞元，然后你發現像“序的”這些詞在中文中基本是不成立的，不過這種Oracle分析器本身就不認識中文，所以只能全部分析，很明顯效率是不好的
- chinese_lexer：這是一種新的中文分析器，前面提到chinese_vgram_lexer這種分析器雖然支持所有的中文字符集，但是效率不高，所以chinese_lexer是對其的改進版本，這種分析器認識很多中文詞匯，能夠比較快查詢，提高效率，不過這種分析器只能支持utf-8字符集
Oracle的全文索引具體可以采用通配符查找、模糊匹配、相關分類、近似查找、條件加權和詞意擴充等方法
環境准備
drop table t purge;
create table t as select * from dba_objects where object_name is not null;
update t set object_name ='高興' where rownum<=2;
commit;
select * from t where object_name like '%高興%';
設置詞法分析器
//設置詞法分析器
BEGIN
ctx_ddl.create_preference ('lexer1', 'chinese_vgram_lexer');
END;
解鎖ctxsys用戶，同時給你的測試賬號(我這里用scott)授權使用ctx_ddl
//解鎖ctxsys用戶同時授權
grant ctxapp to scott;
alter user ctxsys account unlock;
alter user ctxsys identified by ctxsys;
connect ctxsys/ctxsys;
grant execute on ctx_ddl to scott;
connect scott/11;
建立全文索引
//刪除全文索引
drop index idx_content;
//查看數據文件信息
select * from v$datafile;
//建立全文索引
CREATE INDEX idx_content ON t(object_name) indextype is ctxsys.context parameters('lexer lexer1');
注意要點：更新數據時候記得執行全文索引同步命令，否則將看不到更新數據
exec ctx_ddl.sync_index('idx_content','20M');