Date:         Mon, 6 Oct 2008 15:37:21 -0400
Reply-To:     Muthia Kachirayan <muthia.kachirayan@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Muthia Kachirayan <muthia.kachirayan@GMAIL.COM>
Subject:      Re: Multiple observations into multiple variables
In-Reply-To:  <200810061751.m96Al9Wt008886@malibu.cc.uga.edu>
Content-Type: text/plain; charset=ISO-8859-1

On Mon, Oct 6, 2008 at 1:51 PM, Joseph Cormier <jcormier@brconline.com>wrote:

> I have a dataset that looks like this: Each person has multiple injuries...
>
> Year Person Injury Type
> 2005 1 1 A
> 2005 1 2 A
> 2005 1 3 B
> 2005 1 4 B
> 2005 2 1 D
> 2005 2 2 D
> 2005 2 3 E
> 2005 2 4 F
> 2005 2 5 A
> 2004 1 1 B
> 2004 1 2 C
> 2004 1 3 D
>
> I would like to create multiple variables for a single person so that the
> different injuries are aligned in a single observation like this:
>
> Year Person Injury Injury 1  Injury 2  Injury 3
> 2005 1 1 A     0       0
> 2005 1 2 0     A       0
> 2005 1 3 0     0       B
> 2005 1 4 0     0       0
> 2005 2 1 D     0       0
> 2005 2 2 0     D       0
> 2005 2 3 0     0       E
> 2005 2 4 0     0       0
> 2005 2 5 0     0       0
> 2004 1 1 B     0       0
> 2004 1 2 0     C       0
> 2004 1 3 0     0       D
>
> Any help would be greatly appreciated, thanks.
>
>   Joe
>

Without Proc Transpose, it can be done too. The only requirement is to
declare the array

              array inj $1. inj_1 - inj_5;

with inj_1 to inj_5 to match the maximum number of injuries within a PERSON.


data need;
do until (last.person);
   set have;
   by person notsorted;
   array inj $1. inj_1 - inj_5;
   inj[injury] = type;
   output;
   inj[injury] = ' ';
end;
drop injury type;
run;

Muthia Kachirayan