比較長，要有點耐心看，做個實驗。
SQL> create table t_objects as select * from user_objects;
SQL> create table t_tables as select * from user_tables;
SQL> create index ind_objects on T_OBJECTS (object_name);
SQL> create index ind_tables on T_TABLES (table_name);
SQL>exec dbms_stats.gather_table_stats(user,'T_OBJECTS',cascade => true);
SQL>exec dbms_stats.gather_table_stats(user,'T_TABLES',cascade => true);
SQL> alter session set events '10053 trace name context forever,level 1';
SQL> explain plan for select * from t_objects o, t_tables t
2 where t.table_name = o.object_name;
SQL> alter session set events '10053 trace name context off';
一、下面的部分CBO的主要工作是對SQL語句的謂詞進行分析、重寫，把它改爲最符合邏輯的SQL語句
**************************
Predicate Move-Around (PM)
**************************
select * from t_objects o, t_tables t
where t.table_name = o.object_name;
改爲了：
SELECT "O"."OBJECT_NAME" "OBJECT_NAME",
"O"."SUBOBJECT_NAME" "SUBOBJECT_NAME",
"O"."OBJECT_ID" "OBJECT_ID",
"O"."DATA_OBJECT_ID" "DATA_OBJECT_ID",
"O"."OBJECT_TYPE" "OBJECT_TYPE",
"O"."CREATED" "CREATED",
"O"."LAST_DDL_TIME" "LAST_DDL_TIME",
"O"."TIMESTAMP" "TIMESTAMP",
"O"."STATUS" "STATUS",
"O"."TEMPORARY" "TEMPORARY",
"O"."GENERATED" "GENERATED",
"O"."SECONDARY" "SECONDARY",
"T"."TABLE_NAME" "TABLE_NAME",
"T"."TABLESPACE_NAME" "TABLESPACE_NAME",
"T"."CLUSTER_NAME" "CLUSTER_NAME",
"T"."IOT_NAME" "IOT_NAME",
"T"."STATUS" "STATUS",
"T"."PCT_FREE" "PCT_FREE",
"T"."PCT_USED" "PCT_USED",
"T"."INI_TRANS" "INI_TRANS",
"T"."MAX_TRANS" "MAX_TRANS",
"T"."INITIAL_EXTENT" "INITIAL_EXTENT",
"T"."NEXT_EXTENT" "NEXT_EXTENT",
"T"."MIN_EXTENTS" "MIN_EXTENTS",
"T"."MAX_EXTENTS" "MAX_EXTENTS",
"T"."PCT_INCREASE" "PCT_INCREASE",
"T"."FREELISTS" "FREELISTS",
"T"."FREELIST_GROUPS" "FREELIST_GROUPS",
"T"."LOGGING" "LOGGING",
"T"."BACKED_UP" "BACKED_UP",
"T"."NUM_ROWS" "NUM_ROWS",
"T"."BLOCKS" "BLOCKS",
"T"."EMPTY_BLOCKS" "EMPTY_BLOCKS",
"T"."AVG_SPACE" "AVG_SPACE",
"T"."CHAIN_CNT" "CHAIN_CNT",
"T"."AVG_ROW_LEN" "AVG_ROW_LEN",
"T"."AVG_SPACE_FREELIST_BLOCKS" "AVG_SPACE_FREELIST_BLOCKS",
"T"."NUM_FREELIST_BLOCKS" "NUM_FREELIST_BLOCKS",
"T"."DEGREE" "DEGREE",
"T"."INSTANCES" "INSTANCES",
"T"."CACHE" "CACHE",
"T"."TABLE_LOCK" "TABLE_LOCK",
"T"."SAMPLE_SIZE" "SAMPLE_SIZE",
"T"."LAST_ANALYZED" "LAST_ANALYZED",
"T"."PARTITIONED" "PARTITIONED",
"T"."IOT_TYPE" "IOT_TYPE",
"T"."TEMPORARY" "TEMPORARY",
"T"."SECONDARY" "SECONDARY",
"T"."NESTED" "NESTED",
"T"."BUFFER_POOL" "BUFFER_POOL",
"T"."ROW_MOVEMENT" "ROW_MOVEMENT",
"T"."GLOBAL_STATS" "GLOBAL_STATS",
"T"."USER_STATS" "USER_STATS",
"T"."DURATION" "DURATION",
"T"."SKIP_CORRUPT" "SKIP_CORRUPT",
"T"."MONITORING" "MONITORING",
"T"."CLUSTER_OWNER" "CLUSTER_OWNER",
"T"."DEPENDENCIES" "DEPENDENCIES",
"T"."COMPRESSION" "COMPRESSION",
"T"."DROPPED" "DROPPED"
FROM "TEST"."T_OBJECTS" "O", "TEST"."T_TABLES" "T"
WHERE "T"."TABLE_NAME" = "O"."OBJECT_NAME"
AND "T"."TABLE_NAME" ='T_OBJECTS'
二、下面是解釋trace文件用到的一些縮寫的指標定義
Legend
The following abbreviations are used by optimizer trace.
CBQT - cost-based query transformation
JPPD - join predicate push-down
FPD - filter push-down
PM - predicate move-around
CVM - complex view merging
SPJ - select-project-join
SJC - set join conversion
SU - subquery unnesting
OBYE - order by elimination
ST - star transformation
qb - query block
LB - leaf blocks
DK - distinct keys
LB/K - average number of leaf blocks per key
DB/K - average number of data blocks per key
CLUF - clustering factor
NDV - number of distinct values
Resp - response cost
Card - cardinality
Resc - resource cost
NL - nested loops (join)
SM - sort merge (join)
HA - hash (join)
CPUCSPEED - CPU Speed
IOTFRSPEED - I/O transfer speed
IOSEEKTIM - I/O seek time
SREADTIM - average single block read time
MREADTIM - average multiblock read time
MBRC - average multiblock read count
MAXTHR - maximum I/O system throughput
SLAVETHR - average slave I/O throughput
dmeth - distribution method
1: no partitioning required
2: value partitioned
4: right is random (round-robin)
512: left is random (round-robin)
8: broadcast right and partition left
16: broadcast left and partition right
32: partition left using partitioning of right
64: partition right using partitioning of left
128: use hash partitioning dimension
256: use range partitioning dimension
2048: use list partitioning dimension
1024: run the join in serial
0: invalid distribution method
sel - selectivity
ptn – partition
三、綁定變量的描述，由於這裏沒有使用綁定變量，所以沒有信息
*******************************************
Peeked values of the binds in SQL statement
*******************************************
四、一些修復的bug信息
*********************************
Bug Fix Control Environment
*********************************
fix 4611850 = enabled
fix 4663804 = enabled
fix 4663698 = enabled
fix 4545833 = enabled
fix 3499674 = disabled
fix 4584065 = enabled
…………………………………省略…………………………
五、性能相關的初始化參數
*************************************
PARAMETERS WITH DEFAULT VALUES
******************************
optimizer_mode_hinted = false
parallel_execution_enabled = true
optimizer_features_enable = 10.2.0.4
cpu_count = 2
hash_area_size = 131072
bitmap_merge_area_size = 1048576
sort_area_size = 65536
pga_aggregate_target = 198656 KB
_db_file_optimizer_read_count = 16
…………………………………省略…………………………
六、SQL引用對象的基本信息(具體的意思可以看標題二的內容)
下列的信息在CBO做執行計劃代價的計算上都要作爲輸入的值。
***************************************
BASE STATISTICAL INFORMATION
**************************************
Table Stats::
Table: T_TABLES Alias: T
#Rows: 41 #Blks: 5 AvgRowLen: 192.00
Column (#1): TABLE_NAME(VARCHAR2)
AvgLen: 22.00 NDV: 41 Nulls: 0 Density: 0.02439
Index Stats::
Index: IND_TABLES Col#: 1
LVLS: 0 #LB: 1 #DK: 41 LB/K: 1.00 DB/K: 1.00 CLUF: 8.00
***********************
Table Stats::
Table: T_OBJECTS Alias: O
#Rows: 373 #Blks: 8 AvgRowLen: 83.00
Column (#1): OBJECT_NAME(VARCHAR2)
AvgLen: 20.00 NDV: 351 Nulls: 0 Density: 0.0026895
Histogram: HtBal #Bkts: 254 UncompBkts: 254 EndPtVals: 242
Index Stats::
Index: IND_OBJECTS Col#: 1
LVLS: 1 #LB: 2 #DK: 351 LB/K: 1.00 DB/K: 1.00 CLUF: 133.00
************************************************************
七、CBO計算每個對象單獨訪問的代價
Card: Original: 373 原紀錄數 Rounded: 1 輸出紀錄數，可以看到兩張表的訪問索引的訪問方式cost低，所以都走索引。
T_OBJECTS的訪問評估：
SINGLE TABLE ACCESS PATH
-----------------------------------------
BEGIN Single Table Cardinality Estimation
-----------------------------------------
Using density: 0.0026895 of col #1 as selectivity of unpopular value pred
Table: T_OBJECTS Alias: O
Card: Original: 373 Rounded: 1 Computed: 1.00 Non Adjusted: 1.00
-----------------------------------------
END Single Table Cardinality Estimation
-----------------------------------------
Access Path: TableScan
Cost: 3.02 Resp: 3.02 Degree: 0
Cost_io: 3.00 Cost_cpu: 131792
Resp_io: 3.00 Resp_cpu: 131792
Using density: 0.0026895 of col #1 as selectivity of unpopular value pred
Access Path: index (AllEqRange)
Index: IND_OBJECTS
resc_io: 2.00 resc_cpu: 16233
ix_sel: 0.0026895 ix_sel_with_filters: 0.0026895
Cost: 2.00 Resp: 2.00 Degree: 1
Best:: AccessPath: IndexRange Index: IND_OBJECTS
Cost: 2.00 Degree: 1 Resp: 2.00 Card: 1.00 Bytes: 0
T_TABLES的訪問評估：
SINGLE TABLE ACCESS PATH
-----------------------------------------
BEGIN Single Table Cardinality Estimation
-----------------------------------------
Table: T_TABLES Alias: T
Card: Original: 41 Rounded: 1 Computed: 1.00 Non Adjusted: 1.00
-----------------------------------------
END Single Table Cardinality Estimation
-----------------------------------------
Access Path: TableScan
Cost: 3.01 Resp: 3.01 Degree: 0
Cost_io: 3.00 Cost_cpu: 44747
Resp_io: 3.00 Resp_cpu: 44747
Access Path: index (AllEqRange)
Index: IND_TABLES
resc_io: 2.00 resc_cpu: 15533
ix_sel: 0.02439 ix_sel_with_filters: 0.02439
Cost: 2.00 Resp: 2.00 Degree: 1
Best:: AccessPath: IndexRange Index: IND_TABLES
Cost: 2.00 Degree: 1 Resp: 2.00 Card: 1.00 Bytes: 0
***************************************
八、CBO計算列出兩個表關聯方式，並計算出每一種關聯方式的代價，最終選擇最小的cost
***************************************
OPTIMIZER STATISTICS AND COMPUTATIONS
***************************************
GENERAL PLANS
***************************************
Considering cardinality-based initial join order.
Permutations for Starting Table :0
***********************
Join order[1]: T_TABLES[T]#0 T_OBJECTS[O]#1—連接1
***************
Now joining: T_OBJECTS[O]#1
***************
NL Join
Outer table: Card: 1.00 Cost: 2.00 Resp: 2.00 Degree: 1 Bytes: 192
Inner table: T_OBJECTS Alias: O
Access Path: TableScan
NL Join: Cost: 5.03 Resp: 5.03 Degree: 1
Cost_io: 5.00 Cost_cpu: 147324
Resp_io: 5.00 Resp_cpu: 147324
kkofmx: index filter:"T"."TABLE_NAME"="O"."OBJECT_NAME" AND "T"."TABLE_NAME"='T_OBJECTS' AND "O"."OBJECT_NAME"='T_OBJECTS'
Using density: 0.0026895 of col #1 as selectivity of unpopular value pred
Access Path: index (AllEqJoin)
Index: IND_OBJECTS
resc_io: 2.00 resc_cpu: 16233
ix_sel: 0.0026895 ix_sel_with_filters: 0.0026895
NL Join: Cost: 4.01 Resp: 4.01 Degree: 1
Cost_io: 4.00 Cost_cpu: 31766
Resp_io: 4.00 Resp_cpu: 31766
Best NL cost: 4.01
resc: 4.01 resc_io: 4.00 resc_cpu: 31766
resp: 4.01 resp_io: 4.00 resp_cpu: 31766
Join Card: 1.00 = outer (1.00) * inner (1.00) * sel (1)
Join Card - Rounded: 1 Computed: 1.00
SM Join
Outer table:
resc: 2.00 card 1.00 bytes: 192 deg: 1 resp: 2.00
Inner table: T_OBJECTS Alias: O
resc: 2.00 card: 1.00 bytes: 83 deg: 1 resp: 2.00
using dmeth: 2 #groups: 1
SORT resource Sort statistics
Sort width: 231 Area size: 202752 Max Area size: 40684544
Degree: 1
Blocks to Sort: 1 Row size: 102 Total Rows: 1
Initial runs: 1 Merge passes: 0 IO Cost / pass: 0
Total IO sort cost: 0 Total CPU sort cost: 5819699
Total Temp space used: 0
SM join: Resc: 5.01 Resp: 5.01 [multiMatchCost=0.00]
SM cost: 5.01
resc: 5.01 resc_io: 4.00 resc_cpu: 5851465
resp: 5.01 resp_io: 4.00 resp_cpu: 5851465
HA Join
Outer table:
resc: 2.00 card 1.00 bytes: 192 deg: 1 resp: 2.00
Inner table: T_OBJECTS Alias: O
resc: 2.00 card: 1.00 bytes: 83 deg: 1 resp: 2.00
using dmeth: 2 #groups: 1
Cost per ptn: 0.50 #ptns: 1
hash_area: 0 (max=0) Hash join: Resc: 4.51 Resp: 4.51 [multiMatchCost=0.00]
HA cost: 4.51
resc: 4.51 resc_io: 4.00 resc_cpu: 2941866
resp: 4.51 resp_io: 4.00 resp_cpu: 2941866
Best:: JoinMethod: NestedLoop--消耗最少的
Cost: 4.01 Degree: 1 Resp: 4.01 Card: 1.00 Bytes: 275
***********************
Best so far: Table#: 0 cost: 2.0027 card: 1.0000 bytes: 192
Table#: 1 cost: 4.0055 card: 1.0032 bytes: 275
***********************
Join order[2]: T_OBJECTS[O]#1 T_TABLES[T]#0--連接2
***************
Now joining: T_TABLES[T]#0
***************
NL Join
Outer table: Card: 1.00 Cost: 2.00 Resp: 2.00 Degree: 1 Bytes: 83
Inner table: T_TABLES Alias: T
Access Path: TableScan
NL Join: Cost: 5.01 Resp: 5.01 Degree: 1
Cost_io: 5.00 Cost_cpu: 60980
Resp_io: 5.00 Resp_cpu: 60980
kkofmx: index filter:"T"."TABLE_NAME"="O"."OBJECT_NAME" AND "T"."TABLE_NAME"='T_OBJECTS' AND "O"."OBJECT_NAME"='T_OBJECTS'
Access Path: index (AllEqJoin)
Index: IND_TABLES
resc_io: 1.00 resc_cpu: 9261
ix_sel: 0.02439 ix_sel_with_filters: 0.02439
NL Join: Cost: 3.00 Resp: 3.00 Degree: 1
Cost_io: 3.00 Cost_cpu: 25494
Resp_io: 3.00 Resp_cpu: 25494
Best NL cost: 3.00
resc: 3.00 resc_io: 3.00 resc_cpu: 25494
resp: 3.00 resp_io: 3.00 resp_cpu: 25494
Join Card: 1.00 = outer (1.00) * inner (1.00) * sel (1)
Join Card - Rounded: 1 Computed: 1.00
SM Join
Outer table:
resc: 2.00 card 1.00 bytes: 83 deg: 1 resp: 2.00
Inner table: T_TABLES Alias: T
resc: 2.00 card: 1.00 bytes: 192 deg: 1 resp: 2.00
using dmeth: 2 #groups: 1
SORT resource Sort statistics
Sort width: 231 Area size: 202752 Max Area size: 40684544
Degree:1 Blocks to Sort:1 Row size: 222 Total Rows:1
Initial runs: 1 Merge passes: 0 IO Cost / pass: 0
Total IO sort cost: 0 Total CPU sort cost: 5819699
Total Temp space used: 0
SM join: Resc: 5.01 Resp: 5.01 [multiMatchCost=0.00]
SM cost: 5.01
resc: 5.01 resc_io: 4.00 resc_cpu: 5851465
resp: 5.01 resp_io: 4.00 resp_cpu: 5851465
HA Join
Outer table:
resc: 2.00 card 1.00 bytes: 83 deg: 1 resp: 2.00
Inner table: T_TABLES Alias: T
resc: 2.00 card: 1.00 bytes: 192 deg: 1 resp: 2.00
using dmeth: 2 #groups: 1
Cost per ptn: 0.50 #ptns: 1
hash_area: 0 (max=0) Hash join: Resc: 4.51 Resp: 4.51 [multiMatchCost=0.00]
HA cost: 4.51
resc: 4.51 resc_io: 4.00 resc_cpu: 2941866
resp: 4.51 resp_io: 4.00 resp_cpu: 2941866
Best:: JoinMethod: NestedLoop—消耗最少的
Cost: 3.00 Degree: 1 Resp: 3.00 Card: 1.00 Bytes: 275
***********************
Best so far: Table#: 1 cost: 2.0028 card: 1.0032 bytes: 83
Table#: 0 cost: 3.0044 card: 1.0032 bytes: 275
(newjo-stop-1) k:0, spcnt:0, perm:2, maxperm:2000
*********************************
Number of join permutations tried: 2
*********************************
(newjo-save) [1 0 ]
Final - All Rows Plan: Best join order: 2—連接1最小cost爲4.01，連接2最小cost爲3.00，所以最終選擇了連接2。
Cost: 3.0044 Degree: 1 Card: 1.0000 Bytes: 275
Resc: 3.0044 Resc_io: 3.0000 Resc_cpu: 25494
Resp: 3.0044 Resp_io: 3.0000 Resc_cpu: 25494
九、執行計劃
============
Plan Table
============
----------------------------------------------------+-----------------------------------+
| Id | Operation | Name | Rows | Bytes | Cost | Time |
----------------------------------------------------+-----------------------------------+
| 0 | SELECT STATEMENT | | | | 3 | |
| 1 | MERGE JOIN CARTESIAN | | 1 | 275 | 3 | 00:00:01 |
| 2 | TABLE ACCESS BY INDEX ROWID | T_OBJECTS | 1 | 83 | 2 | 00:00:01 |
| 3 | INDEX RANGE SCAN | IND_OBJECTS| 1 | | 1 | 00:00:01 |
| 4 | BUFFER SORT | | 1 | 192 | 1 | 00:00:01 |
| 5 | TABLE ACCESS BY INDEX ROWID | T_TABLES | 1 | 192 | 1 | 00:00:01 |
| 6 | INDEX RANGE SCAN | IND_TABLES | 1 | | 0 | |
----------------------------------------------------+-----------------------------------+
Predicate Information:
----------------------
3 - access("O"."OBJECT_NAME"='T_OBJECTS')
6 - access("T"."TABLE_NAME"='T_OBJECTS')
可以看到最終CBO選擇的是nestloop,這與執行計劃的MERGE JOIN CARTESIAN 對應，笛卡兒算法：是每個集合的任務一個成員都要與其他集合的每個成員進行匹配。