Date:         Sun, 20 Feb 2005 07:15:15 -0500
Reply-To:     "Richard A. DeVenezia" <radevenz@IX.NETCOM.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Richard A. DeVenezia" <radevenz@IX.NETCOM.COM>
Subject:      Re: Newbie Question on how to subset a data set by unique values
              of a variable

enstrophy wrote:
> Hi,
>   I'm new to SAS and am confronted by a task of subsetting
> a data set accroding to unique values of a given variable.
> For example, for the following data set:

Welcome to SAS and SAS-L !

> Label Price
> aa     1.2
> bb     2.4
> aa     3.0
> cc     2.4
> cc     1.5
>
> the objective is to create several data sets, each with
> a unique Label, i.e.,

Often a dubious objective.

>
> data set 1:
> Label Price
> aa     1.2
> aa     3.0
>
> data set 2:
> Label  Price
> bb      2.4
>
>
> data set 3:
> Label  Price
> cc     2.4
> cc     1.5
>
> I wonder what would would be the most straightforward
> way of doing this? I was thinking about using
> proc freq to get the unique values of Label, then
> looping through each value, and subsetting the original
> data set according to it. However, I have not been
> able to figure out how to perform a loop for anything
> other than an array. Any help will be greatly appreciated.
>
> -Yu

Since you are a self proclaimed newbie, take pause.
Typically, breaking a data set into parts is unnecessary.

A BY statement causes SAS to perform its processing per groups of rows of a
table.
You may only need to SORT the data and the REPORT it!

proc sort data=following ;
  BY Label;
run;

filename report temp;
ods html file=report style=sasweb;

proc print data=following;
  BY Label;
run;

ods html close;
--
Richard A. DeVenezia
http://www.devenezia.com/