topic Re: Non-integer index of a vector in SAS/IML Software and Matrix Computations

Non-integer index of a vector

stoffprof — Fri, 02 Aug 2013 15:18:10 GMT

I've noticed that using a non-integer index on a vector returns a value. For example, if I set

x = 1:5;

y = x[{1.2 5.7}];

then y returns a 2-vector {1, 5}.

It seems that the indexes are having the floor function applied before being used to find elements of x. Is this correct?

Re: Non-integer index of a vector

Rick_SAS — Fri, 02 Aug 2013 15:29:43 GMT

My experiment indicates that there is some fuzzing as well, so that x[1-epsilon] still returns x[1]:

x=1:5;

i = 2 - 10##-T(11:16);

y = x;

print i[format=20.16] y;

Re: Non-integer index of a vector

stoffprof — Fri, 02 Aug 2013 15:46:14 GMT

Isn't that what we'd expect from the floor function? Anything less than 2 (even if just epsilon) is 1?

Re: Non-integer index of a vector

Rick_SAS — Fri, 02 Aug 2013 19:14:33 GMT

Add this to the end of the previous program:

f = floor(i);

print i[format=20.16] f y;

You'll see that the conversion of noninteger subscripts is different from a simple floor. The fuzzing of near-integer values is important to correctly handle floating point arithmetic. If you reread my previous post, you'll see that I stated that x[1-epsilon] = x[1], not x[0] (which would be an error).