仅在连续行上按SQL分组

可以使用变量将组ID分配给相邻行:

select t.*,
        (@grp := IF(@sru = CONCAT_WS(':', site, rev, lower), 
                    IF(@sru := CONCAT_WS(':', site, rev, upper), @grp, @grp),
                    IF(@sru := CONCAT_WS(':', site, rev, upper) , @grp + 1, @grp + 1)
                   )
            ) as grp
from (select t.*
      from t
      order by site, size, lower
     ) cross join
     (select @grp := 0, @sru := '') params;

然后您可以聚合以获得所需的内容:

select site, size, min(lower) as lower, max(upper) as upper
from (select t.*,
             (@grp := IF(@sru = CONCAT_WS(':', site, rev, lower), 
                         IF(@sru := CONCAT_WS(':', site, rev, upper), @grp, @grp),
                         IF(@sru := CONCAT_WS(':', site, rev, upper) , @grp + 1, @grp + 1)
                        )
             ) as grp
      from (select t.*
            from t
            order by site, size, lower
           ) cross join
           (select @grp := 0, @sru := '') params
     ) t
group by size, size, grp;

Here 是一个SQL助手。