统计时间分布脚本（mysql）

-- 查询超时记录
select * from visit_record t where t.accesstime>to_date('2013-09-1','yyyy-MM-dd') and t.username is not null and t.result='成功' and t.totaltimecost>5000 and t.clientip <> '192.168.112.53' and t.clientip <> '192.168.112.200' and t.clientip <> '192.168.112.245' order by t.totaltimecost desc ;

-- 导出日期范围内的日志信息（已xml格式导出为stat.xml，使用navicat工具导入进stat表）
select t.username,t.accesstime,t.totaltimecost,t.accessurl from visit_record t where t.accesstime>to_date('2013-09-1','yyyy-MM-dd') and t.username is not null and t.result='成功' and t.clientip <> '192.168.112.53' and t.clientip <> '192.168.112.200' and t.clientip <> '192.168.112.245' 
 and (t.accessurl='/newOA/m_flow/navPage.do' or 
  t.accessurl='/newOA/m_commWorkflow/submitNode_2p0.do' or 
  t.accessurl='/newOA/user/showDaiBanxiangJsp.do' or 
  t.accessurl='/newOA/user/showDaiBanXiangList.do'  )
order by t.totaltimecost desc ;

-- 统计超时率(myql库中)
select 
	t.ACCESSURL as url 
	,COUNT(t.ACCESSURL) as 发生次数
	,COUNT(IF(t.TOTALTIMECOST>5000,t.TOTALTIMECOST,null)) as 超时次数
	,COUNT(IF(t.TOTALTIMECOST>5000,t.TOTALTIMECOST,null))/COUNT(t.ACCESSURL) as 超时率
	,ROUND(avg(t.TOTALTIMECOST)) as 平均耗时
	,max(t.TOTALTIMECOST) as 最大耗时 
				from stat t 
				group by t.ACCESSURL order by 平均耗时 desc;

-- 时间分布（异常统计）

select 
	t.ACCESSURL as url 
	,COUNT(IF(t.TOTALTIMECOST>5000 ,t.TOTALTIMECOST,null))  as 超时次数
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '09:00:00' and '10:00:00'  ,t.TOTALTIMECOST,null)) as '9'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '10:00:00' and '11:00:00'  ,t.TOTALTIMECOST,null)) as '10'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '11:00:00' and '12:00:00'  ,t.TOTALTIMECOST,null)) as '11'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '12:00:00' and '13:00:00'  ,t.TOTALTIMECOST,null)) as '12'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '13:00:00' and '14:00:00'  ,t.TOTALTIMECOST,null)) as '13'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '14:00:00' and '15:00:00'  ,t.TOTALTIMECOST,null)) as '14'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '15:00:00' and '16:00:00'  ,t.TOTALTIMECOST,null)) as '15'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '16:00:00' and '17:00:00'  ,t.TOTALTIMECOST,null)) as '16'
	,COUNT(IF(t.TOTALTIMECOST>5000 and time(t.ACCESSTIME) between '17:00:00' and '18:00:00'  ,t.TOTALTIMECOST,null)) as '17'	
				from stat t 
				group by t.ACCESSURL with ROLLUP;

select 
	t.ACCESSURL as url 
	,COUNT(t.ACCESSURL) as 发生次数
	,COUNT(IF( time(t.ACCESSTIME) between '09:00:00' and '10:00:00'  ,t.TOTALTIMECOST,null)) as '9'
	,COUNT(IF( time(t.ACCESSTIME) between '10:00:00' and '11:00:00'  ,t.TOTALTIMECOST,null)) as '10'
	,COUNT(IF( time(t.ACCESSTIME) between '11:00:00' and '12:00:00'  ,t.TOTALTIMECOST,null)) as '11'
	,COUNT(IF( time(t.ACCESSTIME) between '12:00:00' and '13:00:00'  ,t.TOTALTIMECOST,null)) as '12'
	,COUNT(IF( time(t.ACCESSTIME) between '13:00:00' and '14:00:00'  ,t.TOTALTIMECOST,null)) as '13'
	,COUNT(IF( time(t.ACCESSTIME) between '14:00:00' and '15:00:00'  ,t.TOTALTIMECOST,null)) as '14'
	,COUNT(IF( time(t.ACCESSTIME) between '15:00:00' and '16:00:00'  ,t.TOTALTIMECOST,null)) as '15'
	,COUNT(IF( time(t.ACCESSTIME) between '16:00:00' and '17:00:00'  ,t.TOTALTIMECOST,null)) as '16'
	,COUNT(IF( time(t.ACCESSTIME) between '17:00:00' and '18:00:00'  ,t.TOTALTIMECOST,null)) as '17'
	
				from stat t 
				group by t.ACCESSURL with ROLLUP;

-- 日期分布（正常统计）
select 
	t.ACCESSURL as url 
	,COUNT(t.ACCESSURL) as 发生次数
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=0  ,t.ACCESSTIME,null)) as '星期1'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=1  ,t.ACCESSTIME,null)) as '星期2'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=2  ,t.ACCESSTIME,null)) as '星期3'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=3  ,t.ACCESSTIME,null)) as '星期4'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=4  ,t.ACCESSTIME,null)) as '星期5'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=5  ,t.ACCESSTIME,null)) as '星期6'
	,COUNT(IF( WEEKDAY(t.ACCESSTIME)=6  ,t.ACCESSTIME,null)) as '日'
				from stat t 
				group by t.ACCESSURL 

select 
	t.ACCESSURL as url 
	,COUNT(t.ACCESSURL) as 发生次数
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=0  ,t.ACCESSTIME,null)) as '星期1'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=1  ,t.ACCESSTIME,null)) as '星期2'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=2  ,t.ACCESSTIME,null)) as '星期3'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=3  ,t.ACCESSTIME,null)) as '星期4'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=4  ,t.ACCESSTIME,null)) as '星期5'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=5  ,t.ACCESSTIME,null)) as '星期6'
	,COUNT(IF( t.TOTALTIMECOST>5000 and WEEKDAY(t.ACCESSTIME)=6  ,t.ACCESSTIME,null)) as '日'
				from stat t 
				group by t.ACCESSURL