Date:         Thu, 11 Dec 2008 12:41:54 -0500
Reply-To:     Chang Chung <chang_y_chung@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Chang Chung <chang_y_chung@HOTMAIL.COM>
Subject:      Re: Macro
Comments: To: ganeshkumar.mohanan@GMAIL.COM

On Thu, 11 Dec 2008 08:57:10 -0800, GAN <ganeshkumar.mohanan@GMAIL.COM> wrote:

>What is the difference beteen &a and &&a .?how sas its diffrentiate ?

hi, Ganesh,

Stand alone, they resolve to the same thing. If a part of something larger,
then it may not. See an example below.

%*-- stand alone, &x and &&x are the same. --*;
%symdel a;
%let a = x;
%*-- as it is &a and &&a are the same --*;
%put NOTE: %str(&)a=&a %nrstr(&&)a=&&a;

%*-- as a part of something, they may make differences --*;
%symdel a b xb ay;
%let a = x;
%let b = y;
%let xb = xb;
%let ay = ay;
%put NOTE: %str(&)a%str(&)b=&a&b %nrstr(&&)a%str(&)b=&&a&b;

%*-- on log
NOTE: &a=x &&a=x
NOTE: &a&b=xy &&a&b=ay
--*;

The detailed explanation is a bit more complicated. Given "&&a", Macro
processor takes "&&" and returns "&" first. Then it takes "a" and returns
"a". Thus "&&a" becomes "&a". But sas does not stop here! It then remembers
that it had "&" before and goes back to "&" and "scans forward" and tries to
resolve "&a". And this is not the end of the story, either. What happens
there exists no macro variable a? What if this is also a part of a larger
macro expression? And so on... :-)

This is, in my opinion, not a clean hack. I believe that it is a design
mistake, which "proudly" brought us some ugly compound references in macro like:
  &&var&n, &&&&div&&dept&n, &&&var, or &&&&&&&&something

If you are interested in this topic, then this thread may help:
http://tinyurl.com/6ggpjt or
http://www.listserv.uga.edu/cgi-bin/wa?A2=ind0804B&L=sas-l&P=R17764&D=1&H=0&O=D&T=1

See also Don's best explanation/defense on behalf of sas's:
http://tinyurl.com/6cq4gf or
http://www.listserv.uga.edu/cgi-bin/wa?A2=ind0804B&L=sas-l&D=1&H=0&O=D&T=1&P=28908

cheers,
chang