update a
set a.j_id = b.j_id
from tb a join (select j_name,j_dt,min(j_id) as j_id from tb group by j_name,j_dt)b
   on a.j_name = b.j_name and a.j_dt = b.j_dt

------解决方案--------------------
SQL code

#define OPTTIMES 8
double
polyh2(double a[], double x, unsigned long degree) {
 double result1 = 0.0, result2 = 0.0, result3 = 0.0, result4 = 0.0, xpwr = x;
 double *end = &(a[degree - OPTTIMES + 1]);

 if(degree < OPTTIMES) {
    result1 = *a++;
    end = &(a[degree - 1]);
  while(a<= end) {
    result1 += *a++ * xpwr;
    xpwr *= x;
  }
  return result1;
 }

 result1 = *a++;
 result2 = *a++ * xpwr; xpwr *= x;
 result3 = *a++ * xpwr; xpwr *= x;
 result4 = *a++ * xpwr; xpwr *= x;

 while(a < end) {
    result1 = result1 + *a++ * xpwr;
    xpwr *= x;
    result2 = result2 + *a++ * xpwr;
    xpwr *= x;
    result3 = result3 + *a++ * xpwr;
    xpwr *= x;
    result4 = result4 + *a++ * xpwr;
    xpwr *= x;
    result1 = result1 + *a++ * xpwr;
    xpwr *= x;
    result2 = result2 + *a++ * xpwr;
    xpwr *= x;
    result3 = result3 + *a++ * xpwr;
    xpwr *= x;
    result4 = result4 + *a++ * xpwr;
    xpwr *= x;
 }

 result1 = result1 + result2 + result3 + result4;
 end = &(a[degree - 1]);

 while(a <= end) {
    result1 += *a++ * xpwr;
    xpwr = xpwr * x;
 }
 return result1;
}

------解决方案--------------------
SQL code

create table #test(key_id nvarchar(36),j_name nvarchar(36),j_dt datetime,j_id nvarchar(36))
insert into #test(key_id,j_name,j_dt)
select newid(), 'jack','2011-11-11' union all
select newid(), 'jack','2011-11-11' union all
select newid(), 'jack','2011-11-11' union all
select newid(), 'jack','2011-11-11' union all
select newid(), 'jack','2011-11-11' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack1','2011-11-12' union all
select newid(), 'jack2','2011-11-11' union all
select newid(), 'jack2','2011-11-11' union all
select newid(), 'jack2','2011-11-11' union all
select newid(), 'jack2','2011-11-11' union all
select newid(), 'jack2','2011-11-11' union all
select newid(), 'jack2','2011-11-11'
go
--取每组的第一条的key_id进行j_id的更新
update t1 set j_id=t2.key_id from #test t1 cross apply (select top 1 * from #test where j_name=t1.j_name and j_dt=t1.j_dt) t2

------解决方案--------------------
探讨

引用:

SQL code
create table #test(key_id nvarchar(36),j_name nvarchar(36),j_dt datetime,j_id nvarchar(36))
insert into #test(key_id,j_name,j_dt)