Date:         Thu, 8 Jul 2004 03:28:08 -0700
Reply-To:     Guido <cymraeg_erict@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Guido <cymraeg_erict@HOTMAIL.COM>
Organization: http://groups.google.com
Subject:      R e: Easy data step question
Content-Type: text/plain; charset=ISO-8859-1

Hi Ben,

I expect I've got what you want wrong ...  but try :-

    data sum;
       set test;
       sumapp+apples;
    run;

SUMAPP is implicity retained and initialised to zero though the use of
sum statement.   Personally I would use :-

    data sum;
       set test;
       retain sumapp 0;
       sumapp = sumapp+apples;
    run;

just to make it obvious to other people what is going on.  But you
could use something more interesting like a DoW loop :-

    data sum;
        do until (eof);
           set test end=eof;
           sumapp+apples;
           output;
        end;
    run;

HTH

Guido


-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
ben.powell@CLA.CO.UK
Sent: Thursday, July 08, 2004 11:03 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Easy data step question


Dear SAS-L

What is the best way to create a sum variable in a dataset - apologies
in advance for brain thinking in terms of Excel - without using proc
summary or proc freq? E.g. for variable apples in dataset a, how do I
create variable sum_of_apples which for

(_N_ = 1) = apples_1

and for

(_N_ = 2) = apples_1 + apples_2 ... (_N_ = n) = apples_1 + ... +
apples_n

this being the same as

(where n ^ = 1) sum_of_apples_n = sum_of_apples_(n-1) + apples_n

Sorry to insult your intelligence! I'm trying to disguise my own lack
of
*same* by using the word *best* to obscure the fact I don't have *any*
:-)

Thanks,

Ben.