The SAS/IML language enables you to call other SAS procedures and the DATA step. Since the LEXPERK function does what you need, the simplest solution is to call a DATA step from within your SAS/IML program. For convenience, you can hide the call by encapsulating it within a module. The following module has the same name as the R function that you refer to. It returns the same data, although not necessarily in the same order. The call to permutations(10,4) takes 0.03 seconds on my PC.
proc iml;
/* enumerate permutations of k elements from n items,
including order. By default, the function returns
a matrix with PERM(n, k) rows. Each row is a way to
enumerate k integers from the set {1,2,...,n} */
start permutations(n, k, v=1:n);
submit n k;
data _perk;
keep x1-x&k;
array x[&n] x1-x&n;
n = &n;
k = &k;
do j = 1 to n; x = j; end;
nperm=perm(n, k);
do j=1 to nperm;
call lexperk(j, k, of x);
output;
end;
run;
endsubmit;
use _perk;
read all var _num_ into x;
close _perk;
call delete("_perk");
if IsSkipped(v) then
return( x );
y = shape( v[,x], 0, 2 );
return( y );
finish;
/* Test 1: Return pairs of letters from {a,b,c,d} */
v = "a":"d";
g = permutations(4, 2, v);
print g;
/* Test 2: Time how long it takes to return 4-tuples from 1:10 */
t0 = time();
p = permutations(10, 4);
elapsed = time()-t0;
print (dimension(p))[c={"nrow" "ncol"}];
print elapsed;
