通過v$sqlarea和v$sql視圖查找比較耗費資源的sql

1.從v$sqlarea中找出最多的物理讀的sql

select b.username,
       a.DISK_READS,
       a.EXECUTIONS,
       a.DISK_READS / decode(a.EXECUTIONS, 0, 1, a.EXECUTIONS) ratio,
       a.SQL_TEXT
  from v$sqlarea a, dba_users b
 where a.PARSING_USER_ID = b.user_id
   and a.DISK_READS > 10000
   and b.username not in ('SYS', 'SYSTEM')
 order by a.DISK_READS desc;

 

2.從v$sqlarea中找出最多的邏輯讀的sql

select b.username,
       a.BUFFER_GETS,
       a.EXECUTIONS,
       a.BUFFER_GETS / decode(a.EXECUTIONS, 0, 1, a.EXECUTIONS) ratio,
       a.SQL_TEXT
  from v$sqlarea a, dba_users b
 where a.PARSING_USER_ID = b.user_id
   and a.BUFFER_GETS > 10000
   and b.username not in ('SYS', 'SYSTEM')
 order by a.DISK_READS desc;

 

3.從v$sql中選出最佔用資源的sql

select *
  from (select a.ADDRESS,
               a.SQL_TEXT,
               rank() over(order by a.BUFFER_GETS desc) as rank_bufgets,
               to_char(100 * ratio_to_report(a.BUFFER_GETS) over(), '999.99') as pct_bufgets
          from v$sql a
         where a.PARSING_SCHEMA_NAME = 'PMIS')
 where rank_bufgets < 11

 

4.在第3個sql中可能存在sqltext是一樣的,但是是不同的子cursor的語句,那麼可以查詢v$sql_shared_cursor來查詢爲什麼同一個sqltext的sql會有不同的執行計劃

select * from v$sql_shared_cursor c where c.ADDRESS='&address'