Postgresql表相同的数据最后相邻发生和第一行

我有一个程序，通过每分钟PING来检查网络中计算机的状态。 每次它都会向DB插入一个新行，如下所示（我使用的是postgresql）

我希望结果如下

status   from_time(timestamp)    to_time(timestamp)      id_device(int)

---

如何获得此输出？

这是间隙和孤岛问题。可通过以下方式解决此问题：

select t.status, 
   t.from_time, 
   coalesce(CAST(lead(from_time) over (partition by id_device order by from_time) AS varchar(20)), 'NOW') to_date, 
   t.id_device
from
(
    select t.status, min(checking_time) from_time, t.id_device
    from
    (
        select *, row_number() over (partition by id_device, status order by checking_time) - 
                  row_number() over (partition by id_device order by checking_time) grn
        from data
    ) t
    group by t.id_device, grn, t.status
) t
order by  t.id_device, t.from_time

最关键的是最嵌套的子查询，其中我使用两个

row_number

函数来隔离设备上连续出现的相同状态。一旦有了

grn

值，剩下的就很简单了

结果

status  from_time           to_time             id_device
------------------------------------------------------------
OK      2017-01-01 00:00:00 2017-01-01 00:01:00 1
Failed  2017-01-01 00:01:00 2017-01-01 00:04:00 1
OK      2017-01-01 00:04:00 NOW                 1
OK      2017-01-01 00:00:00 NOW                 2
OK      2017-01-01 00:00:00 NOW                 3

类似问题

---

谢谢@Radim Bača，它的效果很好。

OK       '2017-01-01 00:00:00'   '2017-01-01 00:01:00'   1
Failed   '2017-01-01 00:01:00'   '2017-01-01 00:04:00'   1
OK       '2017-01-01 00:04:00'   NOW                     1

OK       '2017-01-01 00:00:00'   NOW                     2
OK       '2017-01-01 00:00:00'   NOW                     3

select t.status, 
   t.from_time, 
   coalesce(CAST(lead(from_time) over (partition by id_device order by from_time) AS varchar(20)), 'NOW') to_date, 
   t.id_device
from
(
    select t.status, min(checking_time) from_time, t.id_device
    from
    (
        select *, row_number() over (partition by id_device, status order by checking_time) - 
                  row_number() over (partition by id_device order by checking_time) grn
        from data
    ) t
    group by t.id_device, grn, t.status
) t
order by  t.id_device, t.from_time

status  from_time           to_time             id_device
------------------------------------------------------------
OK      2017-01-01 00:00:00 2017-01-01 00:01:00 1
Failed  2017-01-01 00:01:00 2017-01-01 00:04:00 1
OK      2017-01-01 00:04:00 NOW                 1
OK      2017-01-01 00:00:00 NOW                 2
OK      2017-01-01 00:00:00 NOW                 3