Date:         Mon, 28 Jun 2004 10:55:13 -0400
Reply-To:     "Chang Y. Chung" <chang_y_chung@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Chang Y. Chung" <chang_y_chung@HOTMAIL.COM>
Subject:      Re: Independent Random Numbers
Comments: To: Bruce Bradbury <BruceBrad@INAME.COM>

On Mon, 28 Jun 2004 05:50:04 -0700, Bruce Bradbury <BruceBrad@INAME.COM>
wrote:

>I want to generate two independent random uniform numbers in a data
>step. That is, independent of each other and without serial
>correlation.
>
>I see in the SAS online doc how to do this by using different seed
>variables and the call ranuni statement. However, in this case I want
>to use clock-initialised streams for both vectors of random numbers.
>Is there a simple way of doing this?

Hi, Bruce,

I have some codes, which I believe generate 100 observations with two
independent random variables, x1 and x2... but I am not quite sure. :-)

The idea is to use two call ranuni's with different seeds. The thing I am
not sure is that if they are serially correclated or not, since the two
seeds are derived from the same random number stream generated by the
ranuni() function. :-/

Cheers,
Chang

data one;
  seed1 = int(1e4*ranuni(0));
  seed2 = int(1e4*ranuni(0));
  do i = 1 to 100;
    call ranuni(seed1,x1);
    call ranuni(seed2,x2);
    output;
    keep x1 x2;
  end;
  stop;
run;

/* check corr */
proc corr data=one;
  var x1 x2;
run;