通过检查上一行中的值获取最新行

Get latest row by checking a value in previous row

我想获取处于失败状态的最新行作业,但同一作业的前一行应该是成功 .我不想获得不断失败的工作记录。我正在使用 MySQL 版本 5.6.10

来源Tablejob_status:

Job_name start_date     end_date            status
A        8/3/2022 12 PM 8/3/2022 1.30 PM    failed
B        8/3/2022 12 PM 8/3/2022 1.00 PM    failed
C        8/3/2022  3 PM 8/3/2022 3.10 PM    success
B        8/3/2022  3 PM 8/3/2022 3.30 PM    failed
C        8/3/2022  3 PM 8/3/2022 3.20 PM    success
A        8/3/2022  2 PM 8/3/2022 2.10 PM    success
A        8/3/2022  3 PM 8/3/2022 3.20 PM    failed

期望输出:

Job_name start_date     end_date            status
A        8/3/2022  3 PM 8/3/2022 3.20 PM    failed

我正在使用以下查询,不确定如何检查以前的 运行。

select distinct(*) 
from job_status 
where status = 'failed' 
order by start_date desc

就性能而言,这不会赢得任何奖牌,但是...

select *
from t
where status = 'failed'
and coalesce((
    -- previous status
    -- coalesce is needed when there is no previous row
    select status
    from t as x
    where x.job_name = t.job_name and x.start_date < t.start_date
    order by x.start_date desc
    limit 1
), 'success') = 'success'
and not exists (
    -- no next record exists
    select 1
    from t as x
    where x.job_name = t.job_name and x.start_date > t.start_date
)

未经测试,但您应该能够用一个相关的 exists 子句来表达:

select * 
from job_status s
where s.status='failed' 
and exists (
  select * from job_status s2 
    where s2.job_name = s.job_name 
      and s2.status = 'success' 
      and s2.end_date < s.end_date 
      order by s2.end_date desc
      limit 1)
order by start_date desc;

问题的任何解决方案都将取决于一些非常重要的假设,这些假设应该在问题中得到解决:

  1. 状态只能从失败到成功
  2. “最新”是指最大的行 end_date
  3. end_date 属性是 DATETIME 类型
  4. end_date、job_status 和 job_name 的组合是独一无二的

如果其中任何一个不是这种情况,那么 SQL 中的解决方案是不可能的(并且在没有其他信息的情况下使用过程语言仍然非常困难)。

同样,这不是一个有效的函数,尽管它会受益于上面 4 中描述的组合索引的存在,顺序为 job_status、job_name、end_date (否则对于索引来说这将是一个糟糕的选择)....

SELECT c.*
FROM (
    SELECT a.job_name, MAX(b.end_date) AS last_fail
    FROM job_status a
    JOIN job_status b
    ON a.job_name=b.job_name
    WHERE a.status='success'
    AND b.status='failed'
    GROUP BY a.job_name
) ilv 
JOIN job_status c
ON ilv.job_name=c.job_name
AND ilv.last_fail=c.end_date

真的,这只是分组最大问题的一个变体,在这里仍然每周被问一次,并且 its own chapter in the manual

根据数据的分布和行数,更有效的解决方案可能是读取 table 中的每一行,按 job_name、[=38= 排序](降序)和 end_date 然后 return 紧接在每个“成功”之后的条目,其中 job_name 匹配前一行。这应该是可能的......

SELECT job_name, job_status, start_date, end_date
FROM (
  SELECT @prev_status AS prev_status, @prev_job_name AS prev_job_name,
  a.*,
  @prev_status:=a.status AS currstatus,
  @prev_job_name:=a.job_name AS currjobname
  FROM (
   SELECT b.*
   FROM job_status b
   ORDER BY b.job_name, b.status DESC, d.end_date DESC
  ) a
) c
WHERE c.prev_status-'success' 
AND c.job_status='failed'
AND c.prev_job_name=job_name