Tom, I do not quite get what you are trying to do. You have a table that repeats itself (row 1-10 are equal to row 11-20), with no other differences, and you want to sort the two repetitions in a specific order - why not just keep the first half?

To sort that, you can do something like this:

proc format; invalue sexnum 'Male'=-2 'Female'=-1 other=_same_ ; run;

data temp; set pp(obs=10); sort=input(number,sexnum.); run;

proc sort data=temp out=ww; by sort descending goal; run;

If you really want the repetition of the first 10 rows, you can just append HAVE to itself:

proc append data=ww base=ww; run;

Regards, Søren

On Wed, 10 Aug 2011 08:14:47 -0400, Tom Smith <need_sas_help@YAHOO.COM> wrote:

>I have a dataset pp with two variables (number, goal) as below: > > >number goal >------ ---- >Male SL >Female SL >17 SL >18 SL >24 SL >Male PL >Female PL >17 PL >18 PL >24 PL >Male SL >Female SL >17 SL >18 SL >24 SL >Male PL >Female PL >17 PL >18 PL >24 PL > >I need to reorder the values of the variables in following sequencc and >create dataset ww > > >number goal >------ ---- >Male SL >Male PL >Female SL >Female PL >17 SL >17 PL >18 SL >18 PL >24 SL >24 PL >Male SL >Male PL >Female SL >Female PL >17 SL >17 PL >18 SL >18 PL >24 SL >24 PL > >Thanks you so much