11g 等频直方图下sql不走索引扫描-CSDN博客

昨天下午开发人员报告一条非常奇怪的sql，使用该sql的执行计划使用全部扫描，但是查询列是建有索引的，oracle版本 11.2.0.2
该sql格式如下 select * from doc_JUSTIN where do_no ='0069325106';
当时第一反应是该列数据分布不均匀，但是查询后发现该列选择性极佳
SQL> select count(*),count(distinct do_no) from doc_JUSTIN;

COUNT(*) COUNT(DISTINCTDO_NO)
---------- --------------------
14037996 14037996
检查统计信息，表和索引都是最近收集的，且该索引并没有失效，这就比较奇怪了；
使用autotrace查看执行计划
SQL> set autotrace traceonly
SQL> select * from doc_JUSTIN where do_no ='0069325106';

Execution Plan
----------------------------------------------------------
Plan hash value: 1702056455

-----------------------------------------------------------------------------------
| Id | Operation         | Name          | Rows | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT |               | 7047K| 2177M|   175K (1)| 00:35:07 |
|* 1 | TABLE ACCESS FULL| DOC_JUSTIN    | 7047K| 2177M|   175K (1)| 00:35:07 |
-----------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

1 - filter("DO_NO"='0069325106')

Statistics
----------------------------------------------------------
          1 recursive calls
          0 db block gets
     643568 consistent gets
          0 physical reads
          0 redo size
       7434 bytes sent via SQL*Net to client
        519 bytes received via SQL*Net from client
          2 SQL*Net roundtrips to/from client
          0 sorts (memory)
          0 sorts (disk)
1 rows processed
添加hint，强制走索引，虽然consistent gets值下降为5，但是rows却为7047K，和全表扫描的一样，看来应是CBO的问题了
SQL> select /*+ index(doc_JUSTIN,IDX_JUSTIN_2) */ * from doc_JUSTIN where do_no ='0069325106';

Execution Plan
----------------------------------------------------------
Plan hash value: 1340941743

-----------------------------------------------------------------------------------------------
| Id | Operation                   | Name            | Rows | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT            |                 | 7047K| 2177M|   961K (1)| 03:12:19 |
|   1 | TABLE ACCESS BY INDEX ROWID| DOC_JUSTIN      | 7047K| 2177M|   961K (1)| 03:12:19 |
|* 2 |   INDEX RANGE SCAN          | IDX_JUSTIN_2    | 7047K|       | 19027   (1)| 00:03:49 |
-----------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

2 - access("DO_NO"='0069325106')

Statistics
----------------------------------------------------------
          1 recursive calls
          0 db block gets
          5 consistent gets
          0 physical reads
          0 redo size
       7441 bytes sent via SQL*Net to client
        519 bytes received via SQL*Net from client
          2 SQL*Net roundtrips to/from client
          0 sorts (memory)
          0 sorts (disk)
          1 rows processed

使用10053跟踪一把
另：11g后的生成的跟踪文件跟10g变化很大，不再是生成以pid为标志的文件，可以通过查询v$diag_info获取生成的跟踪文件。
以下是10053的摘录，CBO通过计算比较，发现使用全表扫描的cost要比索引扫描小，所以选择全表扫描；
但是density居然为0.5，这是一个非常错误的值，而oracle是如何得来的？
***********************
Table Stats::
Table: DOC_JUSTIN Alias: DOC_JUSTIN
    #Rows: 14094996 #Blks: 644512 AvgRowLen: 324.00 ChainCnt: 0.00
Index Stats::
Index: IDX_JUSTIN_1 Col#: 78
    LVLS: 2 #LB: 33266 #DK: 56 LB/K: 594.00 DB/K: 11395.00 CLUF: 638142.00
Index: IDX_JUSTIN_2 Col#: 2
    LVLS: 2 #LB: 38100 #DK: 14114520 LB/K: 1.00 DB/K: 1.00 CLUF: 1923680.00
Index: IDX_JUSTIN_3 Col#: 45
    LVLS: 0 #LB: 0 #DK: 0 LB/K: 0.00 DB/K: 0.00 CLUF: 0.00
Index: IDX_JUSTIN_4 Col#: 34 5 3
    LVLS: 2 #LB: 46950 #DK: 78 LB/K: 601.00 DB/K: 8181.00 CLUF: 638147.00
Index: PK_DOC_JUSTIN Col#: 1
    LVLS: 2 #LB: 28243 #DK: 14134692 LB/K: 1.00 DB/K: 1.00 CLUF: 668126.00
Access path analysis for DOC_JUSTIN
***************************************
SINGLE TABLE ACCESS PATH
Single Table Cardinality Estimation for DOC_JUSTIN[DOC_JUSTIN]
Column (#2):
    NewDensity:0.500000, OldDensity:0.000000 BktCnt:5548, PopBktCnt:5548, PopValCnt:1, NDV:205
Column (#2): DO_NO(
    AvgLen: 11 NDV: 205 Nulls: 0 Density: 0.500000
    Histogram: Freq #Bkts: 1 UncompBkts: 5548 EndPtVals: 1
Using density: 0.500000 of col #2 as selectivity of unpopular value pred
Table: DOC_JUSTIN Alias: DOC_JUSTIN
    Card: Original: 14094996.000000 Rounded: 7047498 Computed: 7047498.00 Non Adjusted: 7047498.00
Access Path: TableScan
    Cost: 175559.75 Resp: 175559.75 Degree: 0
      Cost_io: 174557.00 Cost_cpu: 20094349137
      Resp_io: 174557.00 Resp_cpu: 20094349137
Using density: 0.500000 of col #2 as selectivity of unpopular value pred
Access Path: index (AllEqRange)
    Index: IDX_JUSTIN_2
    resc_io: 980892.00 resc_cpu: 22017327324
    ix_sel: 0.500000 ix_sel_with_filters: 0.500000
    Cost: 981990.71 Resp: 981990.71 Degree: 1
Best:: AccessPath: TableScan
         Cost: 175559.75 Degree: 1 Resp: 175559.75 Card: 7047498.00 Bytes: 0

Google一把，找到一篇非常好的参考文档
http://www.adellera.it/investigations/11g_newdensity/11gNewDensity.pdf
原来从11.1.0.6开始，当有直方图时，CBO开始使用新算法计算density，不再使用数据字典dba_tab_columns中储存的density值；
该功能由隐含参数_optimizer_enable_density_improvements决定，当为true时开启
当为等频直方图时，newdensity = 0.5/numrows，但是该表有千万条数据，按说newdensity值应该非常小才对，不知为何会计算出值为0.5；
这应该是导致sql选择全表扫描的原因。
解决方案应该很简单，要么去除该列上的直方图信息，要么设置_optimizer_enable_density_improvements为false；
本来是说要今早执行的，但是早上来到后开发说问题已经解决了，他们把该表drop重建，此时表中只有几十万条记录，可以自动选择索引；
登录数据库检查一下，该列上的等频直方图已经消失，失去了一个验证的好机会

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/15480802/viewspace-708513/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/15480802/viewspace-708513/