Date:         Fri, 11 Apr 2008 15:44:19 -0400
Reply-To:     Ya Huang <ya.huang@AMYLIN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ya Huang <ya.huang@AMYLIN.COM>
Subject:      Chance that ranuni generate two identical random numbers?

Hi there,

I wonder what's the chance that ranuni, when given a fixed seed, will
generate 2 identical random number. It seems that ranuni number is
accurate to 14 decimal place. Does it mean that one seed if used more
that 10E14 times will start to repeat the itself?

I tested with one million records, and found no duplicate:

35   data xx;
36   do i=1 to 1E6;
37   r=ranuni(0);
38   output;
39   end;
40   format r best32.;
41   run;

NOTE: The data set WORK.XX has 1000000 observations and 2 variables.
NOTE: DATA statement used:
      real time           0.56 seconds
      cpu time            0.45 seconds


42
43   proc print data=xx (obs=5);
44   run;

NOTE: There were 5 observations read from the data set WORK.XX.
NOTE: PROCEDURE PRINT used:
      real time           0.00 seconds
      cpu time            0.00 seconds


45
46   proc sql;
47   select *
48   from xx
49   group by r
50   having count(*) > 1
51   ;
NOTE: The query requires remerging summary statistics back with the
original data.
NOTE: No rows were selected.

When I tried to do 1E14, it took too long. So I killed it.

Thanks

Ya