Date:         Thu, 9 Aug 2007 07:11:40 -0700
Reply-To:     sounpra@YAHOO.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Song <sounpra@YAHOO.COM>
Organization: http://groups.google.com
Subject:      Re: generate normal data for 2-populations
Comments: To: sas-l@uga.edu
In-Reply-To:  <1186668366.122610.43630@b79g2000hse.googlegroups.com>
Content-Type: text/plain; charset="us-ascii"

On Aug 9, 9:06 am, soun...@yahoo.com wrote:
> Hi All,
>
> Can someone share with me a more efficient way to generate the
> following data structure other than what I've done without PROC IML?
> Group1 is normal (0,1) and Group2 is normal (1, 1).
>
> Thanks!
>
> sim     group     x1
>  1           A        XXX
>  1           A        XXX
>  1           A        XXX
>  1           B        XXX
>  1           B        XXX
>  1           B        XXX
>  2           A        XXX
>  2           A        XXX
>  2           A        XXX
>  2           B        XXX
>  2           B        XXX
>  2           B        XXX
>
> Code:
>
> data group1;
>    seed1 = 123;
>       do sim = 1 to 2;
>             do i = 1 to 3;
>             group = 'A';
>                     x1 = rannor(seed1+sim);
>             output;
>             end;
>        end;
>   keep sim group x1;
> run;
>
> data group2;
>   seed2 = 321;
>       do sim = 1 to 2;
>           do i = 1 to 3;
>          group = 'B';
>                  x1 = 1 + rannor(seed2+sim);
>           output;
>            end;
>        end;
>   keep sim group x1;
> run;
>
> data mydata;
>  set group1 group2;
> run;
>
> proc sort data=mydata;
>  by sim group;
> run;
>
> Thanks!
>
> Song

This was my other solution:

data group1;
   seed1 = 123; seed2 = 321;
      do sim = 1 to 2;
            do i = 1 to 3;
          group = 'A';
                  x1 = rannor(seed1+sim);
          output;
            end;
            do i = 1 to 3;
                    group = 'B';
           x1 = 1 + rannor(seed2+sim);
           output;
             end;
       end;
  keep sim group x1;
run;