Oracle索引扫描算法

SQL> create table t as select * from dba_objects;  
 
Table created.  
 
SQL> create index idx_t on t(object_id);  
 
Index created.  
 
SQL> BEGIN
  2    DBMS_STATS.GATHER_TABLE_STATS(ownname          => 'TEST',
  3                                  tabname          => 'T',
  4                                  estimate_percent => 100,
  5                                  method_opt       => 'for all columns size auto',
  6                                  degree           => DBMS_STATS.AUTO_DEGREE,
  7                                  cascade          => TRUE);
  8  END;
  9  /  
 
SQL> select leaf_blocks,blevel,clustering_factor from dba_indexes where index_name='IDX_T';  
 
LEAF_BLOCKS	BLEVEL CLUSTERING_FACTOR
----------- ---------- -----------------
	165	     1		    1705
 
LEAF_BLOCKS 叶子块 165个
 
BLEVEL  索引高度-1
 
集群因子；
CLUSTERING_FACTOR =1705
 
SQL> select count(distinct dbms_rowid.rowid_block_number(rowid)) from T;
 
COUNT(DISTINCTDBMS_ROWID.ROWID_BLOCK_NUMBER(ROWID))
---------------------------------------------------
					       1057
 
存储在1057个块中
 
SQL> set linesize 200
SQL> select b.num_rows,
       a.num_distinct,
       a.num_nulls,
       utl_raw.cast_to_number(high_value) high_value,
       utl_raw.cast_to_number(low_value) low_value,
       (b.num_rows - a.num_nulls) "NUM_ROWS-NUM_NULLS",
       utl_raw.cast_to_number(high_value) -
       utl_raw.cast_to_number(low_value) "HIGH_VALUE-LOW_VALUE"
  from dba_tab_col_statistics a, dba_tables b
 where a.owner = b.owner
   and a.table_name = b.table_name
   and a.owner = 'TEST'
   and a.table_name = upper('T')
   and a.column_name = 'OBJECT_ID';  2    3    4    5    6    7    8    9   10   11   12   13   14  
 
  NUM_ROWS NUM_DISTINCT  NUM_NULLS HIGH_VALUE  LOW_VALUE NUM_ROWS-NUM_NULLS HIGH_VALUE-LOW_VALUE
---------- ------------ ---------- ---------- ---------- ------------------ --------------------
     74486	  74486 	 0	77616	       2	      74486		   77614
 
SQL> explain plan for select owner from t where object_id<1000;
 
Explained.
 
SQL> select * from table(dbms_xplan.display());
 
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------
 
--------------------------------------------
Plan hash value: 1594971208
 
-------------------------------------------------------------------------------------
| Id  | Operation		    | Name  | Rows  | Bytes | Cost (%CPU)| Time     |
-------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT	    |	    |	958 | 10538 |	 26   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID| T     |	958 | 10538 |	 26   (0)| 00:00:01 |
|*  2 |   INDEX RANGE SCAN	    | IDX_T |	958 |	    |	  4   (0)| 00:00:01 |
-------------------------------------------------------------------------------------
 
Predicate Information (identified by operation id):
 
PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------
 
--------------------------------------------
---------------------------------------------------
 
   2 - access("OBJECT_ID"<1000)
 
14 rows selected.
 
索引扫描首先要定义到叶子块:
 
定位到叶子块 要扫描 多少个块？？？  需要高度-1个块
 
叶子块个数 乘以 选择性
 
定位到叶子块 要扫描 多少个块？？？
 
回表和集群因子有关:
 
选择性(Selectivity) 列唯一键(Distinct_Keys) 与行数(Num_Rows)的比值。
 
这里有个概念叫有效选择性 ，< 的有效选择性为
 
(limit-low_value)/(high_value-low_value)
 
limit 是限制
1000
 
low_value=2
 
1000-2 有可能扫到的值的范围
 
high_value-low_value  表示总共有多少个值:
 
HIGH_VALUE=77616
 
LOW_VALUE=2
 
HIGH_VALUE-LOW_VALUE=77614
 
LEAF_BLOCKS=165
 
索引扫描的计算公式如下：
cost =
 blevel +
 celiling(leaf_blocks *effective index selectivity) +
 celiling(clustering_factor * effective table selectivity)
 
SQL> select 1+ceil(165*(1000-2)/77614)+ceil(1705*(1000-2)/77614) from dual; 
 
1+CEIL(165*(1000-2)/77614)+CEIL(1705*(1000-2)/77614)
----------------------------------------------------
						  26
 
为啥effective table selectivity和effective index selectivity一样？
 
表和索引都包含指定列的数据 两者当然一样