Date:         Fri, 19 Oct 2007 11:59:20 -0400
Reply-To:     Gina Nicolosi <ginanicolosi@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Gina Nicolosi <ginanicolosi@HOTMAIL.COM>
Subject:      Re: simple count and sum
Comments: To: Greg Curson <gscsrc@HOTMAIL.COM>

On Fri, 19 Oct 2007 08:36:20 -0700, Greg Curson <gscsrc@HOTMAIL.COM> wrote:

>Without using Proc tabulate, how can I sum fields base on
>values of VAL1 and VAL2(summing VAL3) then counting
>the different values of VAL2 and adding them as new values(new
>variables to the dataset)
>to the records as VAL 4(being the count) and VAL5(being the sum)
>
>Sample data
>           VAL2      VAL1                       VAL3
>           0002      0501242                 48500.00
>           0002      0501242
>63045.38
>           0003      0501242                   100.00
>           0003      0501242                   118.40
>           0003      0501242                   200.00
>           0003      0501242                   392.00
>
>Results of new dataset
>
>VAL2      VAL1                   VAL4        VAL5
> 0002      0501242                 2          111545.38
> 0003      0501242                 2               810.40

Greg,
If I'm understanding you correctly (and if the value of the 2nd
observation of VAL4 is supposed to be 4 instead of 2), you could do...

data sampledata;
input val2 val1 val3;
cards;
0002 0501242 48500.00
0002 0501242 63045.38
0003 0501242 100.00
0003 0501242 118.40
0003 0501242 200.00
0003 0501242 392.00
;
proc sort; by val1 val2;
proc means;
by val1 val2;
var val3;
output out=results (rename=(_FREQ_=val4) drop=_TYPE_) sum()=val5;
run;

This will produce
val1 val2 val4 val5
501242 2 2 111545.38
501242 3 4 810.4

-gina