Group consecutive rows that are incremented by 1

问题

Here are my datas :

FirstKey   SecondKey    Sequence
A1         B1           10001
A1         B1           10002
A1         B1           10003
A1         B1           10004
A1         B2           10001
A1         B2           10003
A1         B2           10005
A1         B3           10001
A1         B3           10002
A1         B3           10003
A2         B4           20001
A2         B5           20002
A2         B5           20003
A2         B6           20004

I try to output the following result:

FirstKey   SecondKey    StartSequence EndSequence
A1         B1           10001         10004
A1         B2           10001         10001
A1         B2           10003         10003
A1         B2           10005         10005
A1         B3           10001         10003
A2         B4           20001         20001
A2         B5           20002         20003
A2         B6           20004         20004

Sequence is an integer, not a varchar. I want sequences that follow each other with the same key (FirstKey, SecondKey) to appear on the same row. My result is ordered by FirstKey, SecondKey, Sequence Does anyone know a solution to output this in SQL (I'm running with Sybase 12.5)

回答1:

A common solution utilizes a ROW_NUMBER and the fact that there are gaps in your sequence:

select FirstKey,
   SecondKey,
   min(Sequence),
   max(Sequence)
from 
 (
   select FirstKey,
      SecondKey,
      Sequence,
      -- meaningless value, but the same for sequential rows
      Sequence -
      row_number() over (partition by FirstKey, SecondKey order by Sequence) as grp 
   from tab
 ) dt
group by 
   FirstKey,
   SecondKey,
   grp

来源：https://stackoverflow.com/questions/32991863/group-consecutive-rows-that-are-incremented-by-1