DBMS_STATS的用法

收集表的统计信息（表大小写没关系，USER是一个SQL函数）：

HR@ prod> begin

  2  dbms_stats.gather_table_stats

  3  ( ownname => user ,

  4  tabname => 'EMPLOYEES' ) ;

  5  end ;

  6  /

 

PL/SQL procedure successfully completed.

 

HR@ prod> begin

  2  dbms_stats.gather_table_stats

  3  ( ownname => user ,

  4  tabname => 'employees' ) ;

  5  end ;

  6  /

 

收集整个方案的统计信息：

HR@ prod> begin

  2  dbms_stats.gather_schema_stats

  3  ( ownname => user ) ;

  4  end ;

  5  /

 

PL/SQL procedure successfully completed.

 

收集整个方案中统计信息过时的表的统计信息：

HR@ prod> begin

  2  dbms_stats.gather_schema_stats

  3  ( ownname => user ,

  4    options => 'gather stale' ) ;

  5  end ;

  6  /

 

PL/SQL procedure successfully completed.

 

为所有索引列创建直方图：

HR@ prod> begin

  2  dbms_stats.gather_schema_stats

  3  ( ownname => 'hr' ,

  4    method_opt => 'for all indexed columns size auto' ) ;

  5  end ;

  6  /

 

PL/SQL procedure successfully completed.

 

修改默认收集方式，只为数据倾斜分布的索引列创建直方图：

HR@ prod> conn / as sysdba

Connected.

SYS@ prod> begin

  2  dbms_stats.set_database_prefs

  3  ( pname => 'method_opt' ,

  4    pvalue => 'for all indexed columns size skewonly' ) ;

  5  end ;

  6  /

 

PL/SQL procedure successfully completed.

 

删除统计信息

HR@ prod> begin

  2  dbms_stats.delete_schema_stats

  3  ( ownname => user ) ;

  4  end ;

  5  /

 

PL/SQL procedure successfully completed.

 

HR@ prod> select NUM_ROWS from user_tables where table_name = 'EMPLOYEES10' ;

 

  NUM_ROWS

----------