一：what

 1: 聚集因子 是索引資料順序和表順序的一致性程度
 2: 取值範圍為：最小資料塊數，最大為行數，值越小越好！

二：why

 1：聚集因子會影響sql的執行效率
 2：比如表結構，資料量，索引都一樣，生產和測試環境sql執行效率大不一致，
    這種情況下就可能是聚集因子不一致導致的

三：how

drop table t_test;
create table t_test(t_id number(20),t_num number(20));
insert into t_test 
select level,dbms_random.random from
 
 dual connect by level <= 1000000;
create index t_test_idx on t_test(t_id);

drop table t_test2;
create table t_test2(t_id number(20),t_num number(20));
insert into t_test2 
select * from t_test order by dbms_random.random;
create index t_test2_idx on t_test2(t_id);

explain plan for select * from t_test where
 
 t_id between 2000 and 9000;
select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
Plan hash value: 1773029164
--------------------------------------------------------------------------------
| Id  | Operation                   | Name       | Rows  | Bytes | Cost (%CPU)|
--------------------------------------------------------------------------------
 

|   0 | SELECT STATEMENT            |            |  8315 |   211K|    48   (0)|
|   1 |  TABLE ACCESS BY INDEX ROWID| T_TEST     |  8315 |   211K|    48   (0)|
|*  2 |   INDEX RANGE SCAN          | T_TEST_IDX |  8315 |       |    24   (0)|
--------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
   2 - access("T_ID">=2000 AND "T_ID"<=9000)
Note
-----
   - dynamic sampling used for this statement (level=2)
18 rows selected


--執行計劃
explain plan for select * from t_test2 where t_id between 2000 and 9000;
select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
Plan hash value: 1680898452
-----------------------------------------------------------------------------
| Id  | Operation         | Name    | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |         |  6904 |   175K|   668   (3)| 00:00:09 |
|*  1 |  TABLE ACCESS FULL| T_TEST2 |  6904 |   175K|   668   (3)| 00:00:09 |
-----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
   1 - filter("T_ID">=2000 AND "T_ID"<=9000)
Note
-----
   - dynamic sampling used for this statement (level=2)
17 rows selected



--檢視聚集因子
SELECT index_name,table_name,blevel,leaf_blocks,clustering_factor,num_rows
  FROM user_indexes
 WHERE lower(table_name) LIKE 't_test%';

INDEX_NAME                     TABLE_NAME                         BLEVEL LEAF_BLOCKS CLUSTERING_FACTOR   NUM_ROWS
------------------------------ ------------------------------ ---------- ----------- ----------------- ----------
T_TEST_IDX                     T_TEST                                  2        2226              2401    1000000
T_TEST2_IDX                    T_TEST2                                 2        2226            999612    1000000

分析:
1：一個索引範圍掃描獲取資料，另一個全表掃描
2：cost 一個48，另外一個668，效能提升14倍！