没有直方图的执行计划预估结果集行数计算公式
公式摘抄于《催华-基于ORACLE的SQL优化》
转载请注明链接地址：http://blog.csdn.net/launch_225/article/details/25432661
SQL> select * from t1;
N1 V1
---------- --------
1 a
2 b
2 c
4 c
5 d
6 e
6 f
6 f
9 g
10 h
11 i
N1 V1
---------- --------
12 i
12 i
14 j
15 k
16 l
16 m
18 n
18 rows selected.
create index idx_t1_n1 on t1(n1);
1. ====>没有直方图的预估结果集行数
SQL> select count(1) from t1;
COUNT(1)
----------
18
SQL> select count(distinct n1) from t1;
COUNT(DISTINCTN1)
-----------------
13
SQL> select min(n1),max(n1) from t1;
MIN(N1) MAX(N1)
---------- ----------
1 18
TABLE_NAME COLUMN_NAME DENSITY NUM_BUCKETS HISTOGRAM
1 T1 N1 0.0769230769230769 1 NONE
2 T1 V1 0.0714285714285714 1 NONE
1.1=>范围查询
------------------------------------------------------------------------------
(1)目标列大于指定的val,且位于low_value and high_value之间
selectivity=((HIGH_VALUE-VAL)/(HIGH_VALUE-LOW_VALUE))* NULL_ADJUST
NULL_ADJUST=(NUM_ROWS-NUM_NULLS)/NUM_ROWS
SQL> select ((18-5)/(18-1))*1*18 from dual;
((18-5)/(18-1))*1*18
--------------------
13.7647059
1* select * from t1 where n1>5
SQL> /
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 70 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 14 | 70 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 14 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1">5)
(2)目标列小于指定的val,且位于low_value and high_value之间
selectivity=((VAL-LOW_VALUE)/(HIGH_VALUE-LOW_VALUE))* NULL_ADJUST
NULL_ADJUST=(NUM_ROWS-NUM_NULLS)/NUM_ROWS
SQL> select ((5-1)/(18-1))*1*18 from dual;
((5-1)/(18-1))*1*18
-------------------
4.23529412
SQL> select * from t1 where n1<5
2 /
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 4 | 20 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 4 | 20 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 4 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1"<5)
(3)目标列>=指定的val,且位于low_value and high_value之间
selectivity=((HIGH_VALUE-VAL)/(HIGH_VALUE-LOW_VALUE)+1/NUM_DISTINCT)* NULL_ADJUST
NULL_ADJUST=(NUM_ROWS-NUM_NULLS)/NUM_ROWS
SQL> select ((18-5)/(18-1)+1/13)*1*18 from dual;
((18-5)/(18-1)+1/13)*1*18
-------------------------
15.1493213
SQL> select * from t1 where n1>=5;
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 15 | 75 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 15 | 75 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 15 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1">=5)
(4)目标列<=指定的val,且位于low_value and high_value之间
selectivity=((VAL-LOW_VALUE)/(HIGH_VALUE-LOW_VALUE)+1/NUM_DISTINCT)* NULL_ADJUST
NULL_ADJUST=(NUM_ROWS-NUM_NULLS)/NUM_ROWS
SQL> select ((5-1)/(18-1)+1/13)*1*18 from dual;
((5-1)/(18-1)+1/13)*1*18
------------------------
5.6199095
SQL> select * from t1 where n1<=5;
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 6 | 30 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 6 | 30 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 6 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1"<=5)
(5)列位于指定值VAL1和VAL2之间，且VAL1和VAL2在LOW_VALUE和HIGH_VALUE范围
selectivity=((val2-val1)/(HIGH_VALUE-LOW_VALUE)+2/NUM_DISTINCT)* NULL_ADJUST
NULL_ADJUST=(NUM_ROWS-NUM_NULLS)/NUM_ROWS
SQL> select ((10-5)/(18-1)+2/13)*1*18 from dual;
((10-5)/(18-1)+2/13)*1*18
-------------------------
8.06334842
SQL> select * from t1 where n1 between 5 and 10;
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 8 | 40 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 8 | 40 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 8 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1">=5 AND "N1"<=10)
------------------------------------------------------------------------------
1.2 =》等值查询
(1) 等值查询且列上没有NULL值没有直方图的计算公式
selectivity_without_null=(1/num_distinct)
SQL> select 1/13*18 from dual;
1/13*18
----------
1.38461538
SQL> select * from t1 where n1=8;
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 5 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 1 | 5 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 1 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1"=8)
(2)目标列没有直方图且有NULL值(通用于没有直方图时列的选择率的计算)
selectivity_with_null=(1/num_distinct) * ((num_rows-num_nulls)/num_rows)
SQL> select 1/13*((18-0)/18)*18 from dual;
1/13*((18-0)/18)*18
-------------------
1.38461538
--不存在空值也通用
SQL> select * from t1 where n1=8;
Execution Plan
----------------------------------------------------------
Plan hash value: 1577308413
-----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 5 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T1 | 1 | 5 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IDX_T1_N1 | 1 | | 1 (0)| 00:00:01 |
-----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("N1"=8)