sql server 根据执行计划查询耗时操作

 1 with QS as(
 2 select    cp.objtype as object_type,                                                    /*类型*/
 3         db_name(st.dbid) as [database],                                                /*数据库*/
 4         object_schema_name(st.objectid,st.dbid) as [schema],                        /*架构*/
 5         object_name(st.objectid,st.dbid) as [object],                                /*对象名*/
 6         convert(char(16),qs.creation_time,120) as plan_creation,                    /*计划生成时间*/
 7         convert(char(16),qs.last_execution_time,120) as last_execution,                /*最后执行时间*/
 8         qs.plan_generation_num,
 9         qs.execution_count,                                                            /*执行次数*/
10         qs.total_elapsed_time/(1000*qs.execution_count) as avg_elapesd_seconds,        /*总花费时间ms*/
11         qs.total_worker_time/(1000*qs.execution_count) as avg_cpu_cost,                /*平均cpu耗时ms*/
12         qs.total_logical_reads/qs.execution_count as avg_logical_reads,                /*平均逻辑读*/
13         qs.total_logical_writes/qs.execution_count as avg_logical_writes,            /*平均逻辑写*/
14         qs.total_physical_reads/qs.execution_count as avg_physical_reads,            /*平均屋里读*/
15         st.text,                                                                    /*执行文本*/
16         qp.query_plan                                                                /*执行计划*/
17 from sys.dm_exec_query_stats qs
18 join sys.dm_exec_cached_plans cp on cp.plan_handle=qs.plan_handle
19 cross apply sys.dm_exec_sql_text(sql_handle) as st
20 cross apply sys.dm_exec_query_plan(qs.plan_handle) as qp
21 )select top 20 * from QS
22  where text like ‘%%‘
23  --and object_type=‘Proc‘
24  --and avg_logical_reads>200
25  and execution_count>100 /*执行次数*/
26  --and last_execution_time>dateadd(mi,-10,getdate())
27 and last_execution >= ‘2016-05-01 00:00:00.000‘ /*最后执行时间*/
28  order by avg_cpu_cost desc
29