查询低效率的sql语句

查找运行系统里bad sql是一个古老的话题, 我们要根据自己的实际情况来分析。绝不能教条的运用下面介绍的这些方法。使用这些SQL语句时，会对系统表产生分组操作，当然也增大了系统的负载。建议大家在系统启动了一段时间后，在半夜负载较轻的时间定时(例如:一个月)来查一查。一定要具体问题具体分析。
    下面是我收藏的一些查找bad sql的方法:


    select * from (select buffer_gets, sql_text
    from v$sqlarea
    where buffer_gets > 500000
    order by buffer_gets desc) where rownum<=30;


    -- 执行次数多的SQL


    select sql_text,executions from
    (select sql_text,executions from v$sqlarea order by executions desc)
    where rownum<81;


    -- 读硬盘多的SQL


    select sql_text,disk_reads from
    (select sql_text,disk_reads from v$sqlarea order by disk_reads desc)
    where rownum<21;


    -- 排序多的SQL


    select sql_text,sorts from
    (select sql_text,sorts from v$sqlarea order by sorts desc)
    where rownum<21;


    --分析的次数太多，执行的次数太少，要用绑变量的方法来写sql


    set pagesize 600;
    set linesize 120;
    select substr(sql_text,1,80) "sql", count(*), sum(executions) "totexecs"
    from v$sqlarea
    where executions < 5
    group by substr(sql_text,1,80)
    having count(*) > 30
    order by 2;


    -- 游标的观察


    set pages 300;
    select sum(a.value), b.name
    from v$sesstat a, v$statname b
    where a.statistic# = b.statistic#
    and b.name = 'opened cursors current'
    group by b.name;
    select count(0) from v$open_cursor;
    select user_name,sql_text,count(0) from v$open_cursor
    group by user_name,sql_text having count(0)>30;


    --查看当前用户&username执行的SQL


   select sql_text from v$sqltext_with_newlines where (hash_value,address) in
    (select sql_hash_value,sql_address from v$session where username='&username')
    order by address,piece;