Date:         Mon, 10 Dec 2007 11:50:24 -0600
Reply-To:     Yu Zhang <zhangyu05@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Yu Zhang <zhangyu05@GMAIL.COM>
Subject:      Re: Proc Tabulate
In-Reply-To:  <445d9dbe0712100916s3c73b4faw9d56f05217f4dd8@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

* hello,
Actualy, we don't need to create an extra dataset. if we assign 1 to a
variable for each record, then specify it as denominator in PCTSUM. it will
do the job.

HTH

Yu

data* new;

retain ind *1*;

infile datalines dsd;

input result $ system $;

if result='normal' then r1=*.*;

else r1=*1*;

datalines;

normal,cardio

normal,cardio

normal,cardio

normal,cardio

normal,gastro

normal,gastro

normal,neuro

normal,neuro

normal,neuro

normal,neuro

normal,skin

normal,skin

normal,skin

normal,feet

abnormal,cardio

abnormal,gastro

abnormal,gastro

abnormal,gastro

abnormal,neuro

abnormal,resp

abnormal,resp

abnormal,resp

abnormal,resp

abnormal,resp

abnormal,skin

abnormal,skin

abnormal,feet

abnormal,feet

;
*

run*;
*

proc* *tabulate* data=new;

class system;

var r1 ind;

table system,r1=''*(sum='N' pctsum<ind>='%') ;

*run*;


On Dec 10, 2007 11:16 AM, Yu Zhang <zhangyu05@gmail.com> wrote:

> Learned this trick from Howard a while ago and took me almost a hour to
> recall how it works. I took the sample data created by TreeFrog. Thank you
> Tree frog!!!   None of works is done by me.
>
> So here it goes:
>
>