Oracle 反向索引（反轉建索引） 理解

• 一 反向索引

  1.1 反向索引的定義

  • 反向索引作為B-tree索引的一個分支，主要是在創建索引時，針對索引列的索引鍵值進行字節反轉，進而實現分散存放到不同葉子節點塊的目的。

   

  1.2 反向索引針對的問題

  • 使用傳統的B-tree索引，當索引的列是按順序產生時，相應的索引鍵值會基本分布在同一個葉塊中。當用戶對該列進行操作時，難免會發生索引塊的爭用。
  • 使用反向索引，將索引列的鍵值進行反轉，實現順序的鍵值分散到不同的葉塊中，從而減少索引塊的爭用。
  • 例如：鍵值1001、1002、1003，反轉后1001、2001、3001，進而分散到不用的葉子節點塊中。

   

  1.3 反向索引應用場景

  • 索引塊成為熱點塊
  • rac環境
    • rac環境下中多節點訪問訪問數據呈現密集且集中的特點，索引熱塊的產生較高。
    • 在范圍檢索不高的rac環境中使用反向索引可有效提高性能。

   

  1.4 反向索引的優點與缺點

  • 優點：降低索引葉子塊的爭用問題，提升系統性能。
  • 缺點：對於范圍檢索，例如：between，>，<時，反向索引無法引用，進而導致全表掃面的產生，降低系統性能。

   

  1.5 反向索引示例說明

• -- 創建兩張相同結構的表，內部結構及數據均引用scott用戶下的emp表SQL> select count(*) from test01;
  
    COUNT(*)
  ----------
  
  SQL> select count(*) from test02;
  
    COUNT(*)
  ----------
  
  
  
  --針對表TEST01的empno列，添加B-tree索引 
  SQL> create index PK_TEST01 on TEST01(EMPNO);
  Index created.
  
  --針對表TEST02的empno列，添加反向索引
  SQL> create index PK_REV_TEST02 on TEST02(EMPNO) REVERSE;
  Index created.
  
  
  --驗證上面的索引，NORMAL/REV表明為反向索引
  SQL> select TABLE_NAME,INDEX_NAME,INDEX_TYPE from user_indexes where INDEX_NAME like '%TEST%';
  
  TABLE_NAME           INDEX_NAME           INDEX_TYPE
  -------------------- -------------------- --------------------
  TEST01               PK_TEST01            NORMAL
  TEST02               PK_REV_TEST02        NORMAL/REV
  
  
  --打開會話追蹤
  SQL> set autotrace traceonly
  
  
  --相同條件查詢，觀察兩表的執行計划
  SQL> select * from TEST01 where empno=7369;
  Execution Plan
  ----------------------------------------------------------
  Plan hash value: 515586510
  
  -----------------------------------------------------------------------------------------
  | Id  | Operation                   | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
  -----------------------------------------------------------------------------------------
  |   0 | SELECT STATEMENT            |           |     1 |    87 |     2   (0)| 00:00:01 |
  |   1 |  TABLE ACCESS BY INDEX ROWID| TEST01    |     1 |    87 |     2   (0)| 00:00:01 |
  |*  2 |   INDEX RANGE SCAN          | PK_TEST01 |     1 |       |     1   (0)| 00:00:01 |
  -----------------------------------------------------------------------------------------
  
  Predicate Information (identified by operation id):
  ---------------------------------------------------
  - access("EMPNO"=7369)
  
  Note
  -----
     - dynamic sampling used for this statement (level=2)
  
  
  Statistics
  ----------------------------------------------------------
   recursive calls
   db block gets
   consistent gets
   physical reads
   redo size
   bytes sent via SQL*Net to client
   bytes received via SQL*Net from client
   SQL*Net roundtrips to/from client
   sorts (memory)
   sorts (disk)
   rows processed
  
  
  
  
  SQL> select * from TEST02 where empno=7369;
  
  Execution Plan
  ----------------------------------------------------------
  Plan hash value: 1053012716
  
  ---------------------------------------------------------------------------------------------
  | Id  | Operation                   | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
  ---------------------------------------------------------------------------------------------
  |   0 | SELECT STATEMENT            |               |     1 |    87 |     2   (0)| 00:00:01 |
  |   1 |  TABLE ACCESS BY INDEX ROWID| TEST02        |     1 |    87 |     2   (0)| 00:00:01 |
  |*  2 |   INDEX RANGE SCAN          | PK_REV_TEST02 |     1 |       |     1   (0)| 00:00:01 |
  ---------------------------------------------------------------------------------------------
  
  Predicate Information (identified by operation id):
  ---------------------------------------------------
  - access("EMPNO"=7369)
  
  Note
  -----
     - dynamic sampling used for this statement (level=2)
  
  
  Statistics
  ----------------------------------------------------------
   recursive calls
   db block gets
   consistent gets
   physical reads
   redo size
   bytes sent via SQL*Net to client
   bytes received via SQL*Net from client
   SQL*Net roundtrips to/from client
   sorts (memory)
   sorts (disk)
   rows processed
  
  
  
  
  -- 相同范圍條件查詢，觀察兩表的執行計划
  SQL> select * from TEST01 where empno between 7350 and 7500;
  
  Execution Plan
  ----------------------------------------------------------
  Plan hash value: 515586510
  
  -----------------------------------------------------------------------------------------
  | Id  | Operation                   | Name      | Rows  | Bytes | Cost (%CPU)| Time     |
  -----------------------------------------------------------------------------------------
  |   0 | SELECT STATEMENT            |           |     2 |   174 |     2   (0)| 00:00:01 |
  |   1 |  TABLE ACCESS BY INDEX ROWID| TEST01    |     2 |   174 |     2   (0)| 00:00:01 |
  |*  2 |   INDEX RANGE SCAN          | PK_TEST01 |     2 |       |     1   (0)| 00:00:01 |
  -----------------------------------------------------------------------------------------
  
  Predicate Information (identified by operation id):
  ---------------------------------------------------
  - access("EMPNO">=7350 AND "EMPNO"<=7500)
  
  Note
  -----
     - dynamic sampling used for this statement (level=2)
  
  
  Statistics
  ----------------------------------------------------------
   recursive calls
   db block gets
   consistent gets
   physical reads
   redo size
   bytes sent via SQL*Net to client
   bytes received via SQL*Net from client
   SQL*Net roundtrips to/from client
   sorts (memory)
   sorts (disk)
   rows processed
  
  
  
  
  
  SQL> select * from TEST02 where empno between 7350 and 7500;
  
  Execution Plan
  ----------------------------------------------------------
  Plan hash value: 3294238222
  
  ----------------------------------------------------------------------------
  | Id  | Operation         | Name   | Rows  | Bytes | Cost (%CPU)| Time     |
  ----------------------------------------------------------------------------
  |   0 | SELECT STATEMENT  |        |     2 |   174 |     3   (0)| 00:00:01 |
  |*  1 |  TABLE ACCESS FULL| TEST02 |     2 |   174 |     3   (0)| 00:00:01 |
  ----------------------------------------------------------------------------
  
  Predicate Information (identified by operation id):
  ---------------------------------------------------
  - filter("EMPNO">=7350 AND "EMPNO"<=7500)
  
  Note
  -----
     - dynamic sampling used for this statement (level=2)
  
  
  Statistics
  ----------------------------------------------------------
   recursive calls
   db block gets
   consistent gets0  redo size
   bytes sent via SQL*Net to client
   bytes received via SQL*Net from client
   SQL*Net roundtrips to/from client
   sorts (memory)
   sorts (disk)
   rows processed

  通過上面的示例可以看到，當使用between條件進行范圍查詢時，采用反向索引的表，並沒有使用索引，而是采用了全表掃面的方式進行檢索。