Make a dataset option that works for all datasets that adds a variable to the PDV that is essentially the observation number. Add it such that it can be used by the procedure or the data step reading the dataset.
Hypothetical Examples:
data cars (obsnum=M Keep=Make Model N );
do p=1 by 1 until (Last.Make);
set SASHELP.CARS( obsnum=N obs=25);
by Make;
if p<4
then output;
end;
Run;
WORK.CARS:
| Make |
Model |
N |
M |
| Acura |
MDX |
1 |
1 |
| Acura |
RSX Type S 2dr |
2 |
2 |
| Acura |
TSX 4dr |
3 |
3 |
| Audi |
A4 1.8T 4dr |
8 |
4 |
| Audi |
A41.8T convertible 2dr |
9 |
5 |
| Audi |
A4 3.0 4dr |
10 |
6 |
PROC TRANSPOSE DATA=SASHELP.FISH(obsnum=N obs=9)
OUT=LONG_TRANSPOSE
PREFIX=Length
NAME=Source
;
BY Species N;
VAR Length1 - Length3;
RUN;
WORK.LONG_TRANSPOSE:
| Species |
N |
Source |
Length1 |
| Bream |
1 |
Length1 |
23.2 |
| Bream |
1 |
Length2 |
25.4 |
| Bream |
1 |
Length3 |
30 |
| Bream |
2 |
Length1 |
24 |
| Bream |
2 |
Length2 |
26.3 |
| Bream |
2 |
Length3 |
31.2 |
| Bream |
3 |
Length1 |
23.9 |
| Bream |
3 |
Length2 |
26.5 |
| Bream |
3 |
Length3 |
31.1 |
