Oracle-多列統計信息

這個是對動態採樣的一個補充。

人爲的建立一個組合列統計信息，只要謂詞條件中出現組合列某一列的時候，就不在使用以前的計算方式，從而避免CBO執行計劃評估誤差問題。

建立組合列：

dbms_stats.create_extended_stats('SH','T2','(n1,n2)');

收集組合列統計信息：

SQL> exec dbms_stats.gather_table_stats('SH','T2',method_opt=>'for columns (n1,n2) size auto');

 

模擬環境：

仍是拿原先的環境試驗；

 

 

SQL> select count(1) from t2 where n1=3 and n2=3 and c1='a';

 

COUNT(1)

----------

498

返回結果集爲498

SQL> select * from table(dbms_xplan.display_cursor(null,0));

 

PLAN_TABLE_OUTPUT

------------------------------------------------------------------------------------------------------------------------------------------------------

SQL_ID 1tpvtscuhy3b6, child number 1

-------------------------------------

select count(1) from t2 where n1=3 and n2=3 and c1='a'

 

Plan hash value: 4191549303

 

----------------------------------------------------------------------------

| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |

----------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | | | 3 (100)| |

| 1 | SORT AGGREGATE | | 1 | 8 | | |

|* 2 | INDEX RANGE SCAN| IDX_T2 | 25 | 200 | 3 (0)| 00:00:01 |

----------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

2 - access("N1"=3 AND "N2"=3 AND "C1"='a')

 

Note

-----

- cardinality feedback used for this statement

 

查看執行計劃，發現執行計劃返回的rows爲 25，跟實際返回結果集498嚴重誤差。

 

建立組合列：

SQL> set serveroutput on

SQL> declare

cg_name varchar2(30);

begin

cg_name:=sys.dbms_stats.create_extended_stats('SH','T2','(n1,n2)');

end;

/

 

 

SQL> select object_id from user_objects where object_name='T2';

 

OBJECT_ID

----------

87688

 

SQL> select col#,intcol#,name from sys.col$ where obj#=87688;

 

COL# INTCOL# NAME

---------- ---------- ------------------------------------------------------------

1 1 C1

2 2 C2

3 3 N1

4 4 N2

0 5 SYS_STUBZH0IHA7K$KEBJVXO5LOHAS

 

能夠看到，建立組合列實際上就是在原表T2上附件了一個‘SYS_STUBZH0IHA7K$KEBJVXO5LOHAS

’列。

 

SQL> exec dbms_stats.gather_table_stats('SH','T2',method_opt=>'for columns (n1,n2) size auto');

 

SQL> select count(1) from t2 where n1=3 and n2=3 and c1='a';

 

COUNT(1)

----------

498

SQL> /

 

COUNT(1)

----------

498

 

必定要連續執行2次，

SQL> select * from table(dbms_xplan.display_cursor(null,0));

 

PLAN_TABLE_OUTPUT

------------------------------------------------------------------------------------------------------------------------------------------------------

SQL_ID 1ngdx15skn5cm, child number 1

-------------------------------------

select count(1) from t2 where n1=3 and n2=3 and c1='a'

 

Plan hash value: 4191549303

 

----------------------------------------------------------------------------

| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |

----------------------------------------------------------------------------

| 0 | SELECT STATEMENT | | | | 3 (100)| |

| 1 | SORT AGGREGATE | | 1 | 8 | | |

|* 2 | INDEX RANGE SCAN| IDX_T2 | 498 | 3984 | 3 (0)| 00:00:01 |

----------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

2 - access("N1"=3 AND "N2"=3 AND "C1"='a')

 

Note

-----

- cardinality feedback used for this statement

 

查看執行計劃，發現返回的rows 是 498 跟實際返回結果集一致。

得出結論：經過建立多列統計信息，也能夠實現動態採樣的功能。