Using Random-Number Functions and CALL Routines

Seed Values

Random-number functions and CALL routines generate streams of random numbers from an initial starting point, called a seed , that either the user or the computer clock supplies . A seed must be a nonnegative integer with a value less than 2 ³¹ -1 (or 2,147,483,647). If you use a positive seed, you can always replicate the stream of random numbers by using the same DATA step. If you use zero as the seed, the computer clock initializes the stream, and the stream of random numbers is not replicable.

Each random-number function and CALL routine generates pseudo-random numbers from a specific statistical distribution. Every random-number function requires a seed value expressed as an integer constant or a variable that contains the integer constant. Every CALL routine calls a variable that contains the seed value. Additionally, every CALL routine requires a variable that contains the generated random numbers.

The seed variable must be initialized prior to the first execution of the function or CALL routine. After each execution of a function, the current seed is updated internally, but the value of the seed argument remains unchanged. After each iteration of the CALL routine, however, the seed variable contains the current seed in the stream that generates the next random number. With a function, it is not possible to control the seed values, and, therefore, the random numbers after the initialization.

Comparison of Random-Number Functions and CALL Routines

Except for the NORMAL and UNIFORM functions, which are equivalent to the RANNOR and RANUNI functions, respectively, SAS provides a CALL routine that has the same name as each random-number function. Using CALL routines gives you greater control over the seed values. As an illustration of this control over seed values, all the random-number CALL routines show an example of interrupting the stream of random numbers to change the seed value.

With a CALL routine, you can generate multiple streams of random numbers within a single DATA step. If you supply a different seed value to initialize each of the seed variables , the streams of the generated random numbers are computationally independent. With a function, however, you cannot generate more than one stream by supplying multiple seeds within a DATA step. The following two examples illustrate the difference.

Examples

Example 1: Generating Multiple Streams from a CALL Routine

This example uses the CALL RANUNI routine to generate three streams of random numbers from the uniform distribution with ten numbers each. See the results in Output 4.1.

Output 4.1: The CALL Routine Example

 Multiple Streams from a CALL Routine                              1   Obs       Seed_1         Seed_2          Seed_3          X1            X2            X3     1     1394231558      512727191       367385659     0.64924       0.23876       0.17108     2     1921384255     1857602268      1297973981     0.89471       0.86501       0.60442     3      902955627      422181009       188867073     0.42047       0.19659       0.08795     4      440711467      761747298       379789529     0.20522       0.35472       0.17685     5     1044485023     1703172173       591320717     0.48638       0.79310       0.27536     6     2136205611     2077746915       870485645     0.99475       0.96753       0.40535     7     1028417321     1800207034      1916469763     0.47889       0.83829       0.89243     8     1163276804      473335603       753297438     0.54169       0.22041       0.35078     9      176629027     1114889939      2089210809     0.08225       0.51916       0.97286    10     1587189112      399894790       284959446     0.73909       0.18622       0.13269 

 

 options nodate pageno=1 linesize=80 pagesize=60;  data multiple(drop=i);     retain Seed_1 1298573062 Seed_2 447801538            Seed_3 631280;     do i=1 to 10;        call ranuni (Seed_1,X1);        call ranuni (Seed_2,X2);        call ranuni (Seed_3,X3);        output;     end;  run;  proc print data=multiple;     title 'Multiple Streams from a CALL Routine';  run; 

Example 2: Assigning Values from a Single Stream to Multiple Variables

Using the same three seeds that were used in Example 1, this example uses a function to create three variables. The results that are produced are different from those in Example 1 because the values of all three variables are generated by the first seed. When you use an individual function more than once in a DATA step, the function accepts only the first seed value that you supply and ignores the rest.

 options nodate pageno=1 linesize=80 pagesize=60;  data single(drop=i);     do i=1 to 3;        Y1=ranuni(1298573062);        Y2=ranuni(447801538);        Y3=ranuni(631280);        output;     end;  run;  proc print data=single;     title 'A Single Stream across Multiple Variables';  run; 

Output 4.2 shows the results. The values of Y1, Y2, and Y3 in this example come from the same random-number stream generated from the first seed. You can see this by comparing the values by observation across these three variables with the values of X1 in Output 4.1.

Output 4.2: The Function Example

 A Single Stream across Multiple Variables                         1     Obs        Y1            Y2            Y3          1      0.64924       0.89471       0.42047      2      0.20522       0.48638       0.99475      3      0.47889       0.54169       0.08225