Date:         Fri, 6 Oct 2006 22:29:43 +0000
Reply-To:     toby dunn <tobydunn@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         toby dunn <tobydunn@HOTMAIL.COM>
Subject:      Re: Compute macro variables
Comments: To: mahesh.tamboli@GMAIL.COM
In-Reply-To:  <1160171628.021691.73030@h48g2000cwc.googlegroups.com>
Content-Type: text/plain; format=flowed

I just wrote a paper for NESUG and SESUG over this very topic:

Go here:

http://www.albany.edu/faculty/msz03/nesug06/cc/cc21.pdf




Toby Dunn

When everything is coming at you all at once, your in the wrong lane.

A truly happy person is someone who can smile and enjoy the scenery on a
detour.





From: mahesh.tamboli@GMAIL.COM
Reply-To: mahesh.tamboli@GMAIL.COM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Compute macro variables
Date: Fri, 6 Oct 2006 14:53:48 -0700

Hi there,

The code below does the following. Based on the current month computes
the year and month for the last 12 months and pulls the relevant data .
The code works, but the issue is the way it is coded. Since 2 new macro
variables need to be computed based on the current month date, a data
_NULL_ step is used. Here, I am trying to explore any other
alternatives that will help me to avoid using a data step (and
work.cust_data dataset).

Appreciate any suggestions..

Thanks,

/* Code Begins here */

%let cm_dt = '1sep06'd;

%macro auto;

%DO N = 1 %TO 12;

   data _NULL_;
      set work.cust_data;      /* Uses a existing temp dataset */

    format year $CHAR4.;
    format month $CHAR2.;

     year  = substr(put(intnx('month', &cm_dt, &N - 8), yymmddn8.),
1,4);
     month = left(substr(put(intnx('month', &cm_dt, &N - 8), yymmddn8.),
5,2);
     call symput('year', year);
     call symput('month',month);

run;

libname old_ds "/marketing/data/&year/&month";

data  result_&N;
    merge old_ds.acct_info (in = key1)
              work.temp11      (in = key2);
   if (key1 and key2)
run;

%END;
%mend;

%auto;   /* Macro Call */