序列重复表重新编号

Question

我想给重复的recods提供序列编号并计算时间延迟而不能这样做

示例：

id time 

A  12h30
A  13h45
A  19h48
B  8h00
B  10h12
C  22h50
D  9h02  
D  16h46

预期查询结果

id time   sequence lag

A  12h30      1    0h00
A  13h45      2    1h15    
A  19h48      3    6h02
B  8h00       1    0h00
B  10h12      2    2h12
C  22h50      1    0h00
D  9h02       1    0h00
D  16h46      2    6h44

我所能做的就是用这个查询重新编号一个连续的序列

SELECT count(*)as rank, a1.id


FROM  table a1, table a2  

where a1.id >=a2.id

group by a1
order by rank, a1.id

最好的做法是什么？

非常感谢

Answer 1

您可以使用row_number()作为序列，lag()计算前一行之间的时差

select 
    id, 
    time,
    row_number() over(partition by id order by time) sequence,
    (case when (lag(id) over (order by id, time)) = id
         then time - (lag(time) over (order by id, time))
         else 0
    end) diff
from mytable t1
order by id, time

您可能必须根据time列的类型修改case语句。