需要修改查询而不使用交叉/外部应用答案

【问题标题】：need to modify the query without using cross/outer apply需要修改查询而不使用交叉/外部应用
【发布时间】：2019-05-24 06:26:58
【问题描述】：

create table PrinterData(
    Dt DateTime not null,
    Status nvarchar(50) not null    
)

insert into PrinterData
values
('05.24.2019 00:00:00', 'READY'),
('05.24.2019 00:01:00', 'READY'),
('05.24.2019 00:02:00', 'NO PAPER'),
('05.24.2019 00:03:00', 'UNKNOWN'),
('05.24.2019 00:04:00', 'UNKNOWN'),
('05.24.2019 00:05:00', 'READY'),
('05.24.2019 00:06:00', 'READY'),
('05.24.2019 00:07:00', 'READY'),
('05.24.2019 00:08:00', 'NO PAPER'),
('05.24.2019 00:09:00', 'UNKNOWN'),
('05.24.2019 00:10:00', 'HAS PAPER'),
('05.24.2019 00:11:00', 'NO PAPER')

我需要知道READY状态和下一个READY或HAS PAPER之间的时间间隔，除了UNKNOWN地位。未知我需要跳过。结果应如下所示：

这是我的查询，但我不喜欢使用交叉/外部应用。在性能和速度方面，我需要您的建议，如何修改此查询。

select a.Dt, a.Status, b.Dt as Dt_Next_Status, b.Status as Next_Status, DATEDIFF(SECOND, a.Dt, b.Dt) as Diff from PrinterData a
outer apply 
(
    select top 1  Dt, Status from PrinterData where Dt > a.Dt and Status not in ('UNKNOWN', 'NO PAPER') 
) b
where a.Status = 'NO PAPER'

【问题讨论】：

标签： sql sql-server tsql sql-server-2012

【解决方案1】：

使用lead()函数

DEMO

select *,datediff(SECOND,dt,nexttime) as diff from 
(
select *,lead(dt) over(order by dt) as nexttime,lead(status) over(order by dt) as nextstatus
from PrinterData
where Status<>'UNKNOWN'
)A where status='NO PAPER'

输出：

Dt                  Status      nexttime            nextstatus  diff
24/05/2019 00:02:00 NO PAPER    24/05/2019 00:05:00 READY       180
24/05/2019 00:08:00 NO PAPER    24/05/2019 00:10:00 HAS PAPER   120
24/05/2019 00:11:00 NO PAPER

【讨论】：

【解决方案2】：

您可以使用左连接和 row_number() 来做到这一点

with cte as
(
select a.Dt, a.Status, b.Dt as Dt_Next_Status, b.Status as Next_Status,
DATEDIFF(SECOND, a.Dt, b.Dt) as Diff,
row_number()over(partition by a.DT order by b.Dt) rn
from PrinterData a
left join
(
    select  Dt, Status from PrinterData where Status not in ('UNKNOWN', 'NO PAPER') 
) b on  b.Dt > a.Dt
where a.Status = 'NO PAPER' 
) 
select * from cte where rn=1

demo in fiddle

【讨论】：