Date:         Fri, 16 May 1997 08:41:14 -0400
Reply-To:     Ed Heaton <edheaton@FENIX2.DOL-ESA.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Ed Heaton <edheaton@FENIX2.DOL-ESA.GOV>
Subject:      PROC SUMMARY: CLASS with NWAY vs. BY

Greetings all;
 
        I have a dataset and I wish to sum certain variables (d, e, and f) based
on the levels of other variables (a, b, and c).  All these variables are
numeric.  I see two ways to accomplish this task with PROC SUMMARY:
 
   PROC SUMMARY
      DATA= sample (KEEP= a b c d e f)
      NWAY
      ;
      CLASS  a b c ;
      VAR d e f ;
      OUTPUT
         OUT= sampout1
         SUM=
         ;
   RUN ;
 
and
 
   PROC SUMMARY
      DATA= sample (KEEP= a b c d e f)
      ;
      BY a b c ;
      VAR d e f ;
      OUTPUT
         OUT= sampout2
         SUM=
         ;
   RUN ;
 
What is the difference?
 
Thanks, in advance, for your insight.
 
Edward Heaton, Systems Development Specialist,
CDSI Information Technology Solutions Company
(A Division of Computer Data Systems, Inc.)
 
On contract for:
U.S. Department of Labor, Room N-4420,
200 Constitution Avenue, Northwest, Washington, DC 20210
(202) 219-7950                                  edheaton@fenix2.dol-esa.gov