我的想法是：先得出唯一的时间列，赋予编号unit和分组grp(二者取值相同) ，如

grp unit t y

1 1 9:22 43
2 2 9:27 43
3 3 9:28 43
4 4 9:29 42.7
5 5 9:30 42.33
6 6 9:31 42.5

再和源数据作合并，调用rank程序，对unit顺序编秩，by grp即可。预想的结果如下

t y unit grp rank

9:22 43.00 1 1 1
9:27 43.00 2 2 1
9:27 43.00 2 2 2
9:27 42.99 2 2 3
9:28 43.00 3 3 1
9:29 42.32 4 4 1
9:29 42.70 4 4 2
9:29 42.50 4 4 3
9:30 42.30 5 5 1
9:30 42.33 5 5 2
9:31 42.50 6 6 1

还可以进一步作首列和最后一列的字符串合并

data a;
input x $ y;
datalines;
9:28 43
9:27 43
9:27 43
9:27 42.99
9:29 42.7
9:29 42.32
9:29 42.5
9:30 42.33
9:30 42.3
9:31 42.5
;
run;

proc sort;
by x;
data b;
set a;
retain id;
by x;
if first.x then id=1;else id=id+1;
x=compress(x||id);
run;

估计你漏了一句

data a;
input x $ y;
datalines;
9:28 43
9:27 43
9:27 43
9:27 42.99
9:29 42.7
9:29 42.32
9:29 42.5
9:30 42.33
9:30 42.3
9:31 42.5
;
run;

proc sort;
by x;
data b;
set a;
retain id;
by x;
if first.x then id=1;else id=id+1;
if first.x+last.x=2 then id=.;
x=compress(x||id);drop id;
proc print;
run;

放在iml里也很简单的!