How to change order of observations in dataset without using proc procedures.

[Sidhu]

How to change order of observations in dataset without using proc procedures like proc sort, proc sql etc.

What I want to ask ... is there any way to shuffle the dataset. As this is question I faced by interview.

For example :

In one dataset :

Subid Invid Sex Age

0001 101   Male 30

0002 202   Male 40

0003 303   Male 50

0004 404  Female 60

The output dataset should be like.

Subjid Invid Sex Age

0004 404 Female 60

0001 101 Male    30

0003 303 Male    50

0002 202 Male    40

[Tom]

Stupid question.

To get that exact result you could use WHERE clauses on SET statements.

data want ;

  set have (where=(sex='Female'))

      have (where=(sex='Male'))

  ;

run;

If they wanted to know if you knew how to use dataset options they should have just asked a question about that.

[PGStats]

Agreed, but you get order subid=4,1,2,3 and 4,1,3,2 was requested. Maybe they wanted

data want;

do i = 1 to ceil(last/2);

     p = last - i + 1;

     set have point=p nobs=last;

     output;

     if i < p then do;

          set have point=i;

          output;

          end;

     end;

stop;

drop i;

run;

PG