SQL Server 2008 R2 sys.dm_exec_sql_text 问题

Question

如果我在我们的生产服务器上运行以下查询：

SELECT DISTINCT TOP 10
t.TEXT QueryName,
s.execution_count AS ExecutionCount,
s.max_elapsed_time / 100000 AS MaxElapsedTime,
ISNULL(s.total_elapsed_time / s.execution_count, 0) / 100000 AS AvgElapsedTime,
s.creation_time AS LogCreatedOn,
ISNULL(s.execution_count / DATEDIFF(s, s.creation_time, GETDATE()), 0) AS FrequencyPerSec
,s.plan_handle, s.plan_generation_num
FROM sys.dm_exec_query_stats s
CROSS APPLY sys.dm_exec_sql_text( s.sql_handle ) t
ORDER BY
s.max_elapsed_time / 100000 DESC

我的主要查询是“创建过程.......”这是否意味着 sql server 重新编译这个存储过程不止一次？

Answer 1

不，关于您的查询，这意味着此程序比其他程序花费更多时间

和

sql server 定期重新编译过程或其内部语句，这是一种预期行为，取决于服务器活动、内存压力、用户命令、RECOMPILE 选项、基础数据更改量等等。

更新：

查询是从 CREATE PROCEDURE 开始的，因为 sql server 以这种方式呈现 SP 的主体。

您甚至可以通过此查询深入了解过程的内部语句：

Select 
    s3.name as [Obj Name], 
    s3.type as [Obj Type], 
    (select top 1 substring(text,(s1.statement_start_offset+2)/2,
    (CASE when s1.statement_end_offset = -1 then len(convert(nvarchar(max),text))*2
    else s1.statement_end_offset end - s1.statement_start_offset) /2 ) FROM sys.dm_exec_sql_text(s1.sql_handle)) as [SQL Statement], 
    execution_count, 
    plan_generation_num, 
    last_execution_time, 
    ((total_worker_time+0.0)/execution_count)/1000 as [avg_worker_time], 
    total_worker_time/1000.0 total_worker_time, 
    last_worker_time/1000.0 last_worker_time, 
    min_worker_time/1000.0 min_worker_time, 
    max_worker_time/1000.0 max_worker_time, 
    ((total_logical_reads+0.0)/execution_count) as [avg_logical_reads], 
    total_logical_reads+0.0 total_logical_reads,
    last_logical_reads+0.0 last_logical_reads, 
    min_logical_reads+0.0 min_logical_reads, 
    max_logical_reads+0.0 max_logical_reads, 
    ((total_logical_writes+0.0)/execution_count) as [avg_logical_writes], 
    total_logical_writes+0.0 total_logical_writes, 
    last_logical_writes+0.0 last_logical_writes, 
    min_logical_writes+0.0 min_logical_writes, 
    max_logical_writes+0.0 max_logical_writes, 
    ((total_logical_writes+0.0)/execution_count + (total_logical_reads+0.0)/execution_count) as [avg_logical_IO], 
    total_logical_writes + total_logical_reads+0.0 total_logical_IO, 
    last_logical_writes +last_logical_reads+0.0 last_logical_IO, 
    min_logical_writes +min_logical_reads+0.0 min_logical_IO, 
    max_logical_writes + max_logical_reads+0.0 max_logical_IO
from sys.dm_exec_query_stats s1
cross apply sys.dm_exec_sql_text(sql_handle) as s2
join sys.objects s3 on ( s2.objectid = s3.object_id )
left join sys.schemas sch on(s3.schema_id = sch.schema_id)
where s2.dbid = db_id()
order by s3.name, s1.sql_handle