我正在尝试创建一个将带来JOB_ID的最后一个STATUS的查询,这在SQL Server 2008中使用rank over partition函数很简单,但在SQL Server 2000上这个函数不可用。
我已设法通过Job_ID升序和DATETIME下降到临时表( #JOB_HISTORY -note,使用:IDENTITY(int, 1,1)
在临时表中生成pk_ID)按顺序放置作业历史记录,但是这里是我卡住的部分,我不知道如何对这张表中的记录进行排名。
SELECT
h1.pk_ID,
h1.Job_ID,
h1.Status,
h1.DATETIME
FROM #JOB_HISTORY h1
ORDER BY h1.pk_ID ASC;
#JOB_HISTORY结果:
pk_ID, Job_ID, STATUS, DATETIME
1 1234 Succeded. 2015-03-30 12:10
2 1234 Failed. 2015-03-30 01:00
3 1234 Failed. 2015-03-28 01:00
4 5678 Failed. 2015-04-02 04:00
5 5678 Succeded. 2015-04-01 04:00
6 5678 Succeded. 2015-03-31 04:00
期望的输出:
Job_ID STATUS DATETIME
1234 Succeded. 2015-03-30 12:10
5678 Failed. 2015-04-02 04:00
答案 0 :(得分:0)
SQL Server可以自由地按照任何顺序插入记录,因此我不会信任IDENTITY(int, 1,1)
总是为您提供正确的序列。它可能在某些版本中有效。
您可以在内部加入子查询,在该查询中获取最大日期时间(或最小值pk_ID)以获取第一行/最后一行。这是获取SQL Server 2000中每个作业的最后结果的工作示例:
use msdb;
select
sysjobhistory.job_id as job_id
-- https://msdn.microsoft.com/en-us/library/ms174997.aspx
, case sysjobhistory.run_status
when 0 then 'Failed'
when 1 then 'Succeeded'
when 2 then 'Retry'
when 3 then 'Canceled'
else cast(sysjobhistory.run_status as varchar(10))
end as run_status
, jobLastHistory.lastDateTime as execution_time
FROM dbo.sysjobhistory as sysjobhistory
inner join (
--last datetime for each job_id
SELECT
history.job_id
, max( --newest
CONVERT(DATETIME, STUFF(STUFF(STUFF( --yyyymmddhhmiss --> datetime
cast(history.run_date as varchar(8))--yyyymmdd
+right('000000'+cast(history.run_time as varchar(8)),6) --hhmiss
,13,0,':'),11,0,':'),9,0,' '))
) AS lastDateTime
FROM dbo.sysjobhistory as history
where history.step_id = 0 --job outcome
group by
history.job_id
) as jobLastHistory
on jobLastHistory.job_id
= sysjobhistory.job_id --same job
and convert(varchar(8),jobLastHistory.lastDateTime,112)
= sysjobhistory.run_date --latest date (yyyymmdd)
and replace(convert(varchar(8),jobLastHistory.lastDateTime,108),':','')
= sysjobhistory.run_time --latest time (hhmiss)
where sysjobhistory.step_id = 0 --job outcome
;