Date:         Mon, 24 Apr 2000 18:31:21 GMT
Reply-To:     "John M. Wildenthal" <jmwildenthal@MY-DEJA.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "John M. Wildenthal" <jmwildenthal@MY-DEJA.COM>
Organization: Deja.com - Before you buy.
Subject:      Re: Macro Question

In article <000001bfae0a$ab109580$0b01a8c0@CORE>,
  "J. Das" <jdas@INTEGRAINFO.COM> wrote:
> Hello All, I have the following problem. Can someone please help me?
>
> I need to perform a Seemingly Unrelated regression (SUR) using PROC
> SYSLIN; But, I need to perform this SUR for many different group of
> industries. In this example, the first group consists of industries
> 0400 and 0600. The next group consists of industries 1711, 1731 and
> 1798.
>
> I need to generate the following code:
>
> PROC SYSLIN DATA=test SUR;
>    MODEL y0400=x0400;
>    MODEL y0600=x0600;
> run;
> PROC SYSLIN DATA=test SUR;
>    MODEL y1711=x1711;
>    MODEL y1731=x1731;
>    MODEL y1798=x1798;
> run;
>
> The problem is I need to perform this PROC SYSLIN for many different
> groups (actually 30 groups).  I need to generate the texts using a
> macro. However, notice that the number of model statements within the
> PROC might vary. How can I generate the above code using a macro.
> Can it be done?
> I'm quite new in SAS/Macro. So, your help will be highly appreciated.
> Thank you very much.

Primitive, but should work:

%LET model1 = 0400 0600 ;
%LET model2 = 1711 1731 1798;
%LET nummods = 2;

%MACRO loopthru;

  %LOCAL i position token ;
  %DO i = 1 %TO &nummods ;

PROC SYSLIN DATA=test SUR;

    %LET position = 1;
    %LET token = %SCAN(&&model&i,&position)
    %DO WHILE %LENGTH(&token) ;

  MODEL y&token = x&token ;

      %LET position = %EVAL(&position+1);
      %LET token = %SCAN(&&model&i,&position)
    %END;

RUN;

  %END;
%MEND;

%loopthru


Sent via Deja.com http://www.deja.com/
Before you buy.