Date:         Thu, 15 Jan 2004 16:55:58 -0500
Reply-To:     Don Stanley <don_stanley@PARADISE.NET.NZ>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Don Stanley <don_stanley@PARADISE.NET.NZ>
Subject:      Re: How to make them into macro vars?

I think Andrea is saying that those 5-9 variables must exist in any created
dataset, according to the code she has created. The PDV is not of concern,
as only one obs is ever in their at a time. I would simply drop the 5
fields and then no issues relating to those 5-9 fields impact on dataset
size exist.

As regards the other issue, which is how to easily modify the code to cope
with more avg variables, I would remove the current macro code, and replace
with this (I'm assuming that the structure of the IF statements applies
throughout)


filename scrinput "c:\two.txt";
filename avgs "c:\one.txt";

%macro model(number=);
 if _n_=1 then do;
   infile avgs end=eof ;
   retain avg1 - avg&number 0 ;
   input avg1 - avg&number ;
   drop avg: ; /* removes the avg variables from any output dataset */
 end ;

 infile scrinput;
 input x ;

 %do i=1 %to %eval(&number - 1 );
   if x < %eval(&i*2) then y = avg&i * x ;
   else
 %end ;

 y= avg&number * x;
%mend;

 options mprint ; /* just to see the generated code */
 data temp;
 %model(number=5)
 run;

There are many ways to do this using arrays, multiple datasteps and so on,
this is near the original code (so I assume it will assist in understanding
it). All Andrea need do if she has say 9 avg variables is change the %model
(number=5) to %model(number=9) etc.

Cheers
Don