Date:         Wed, 3 Dec 2008 16:56:27 -0500
Reply-To:     Akshaya Nathilvar <akshaya.nathilvar@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Akshaya Nathilvar <akshaya.nathilvar@GMAIL.COM>
Subject:      Re: Proc Transpose - long to wide
Comments: To: sas 9 bi user <sas9bi@gmail.com>
In-Reply-To:  <af7d406c0812031346k70472f47k577fb07d79f99d01@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

Proc transpose and rename option will help:
Proc transpose data=have out=want(drop=_name_ rename=(col1= col2= col3=));
    by id seq;
    var response;
Run;

Akshaya

On Wed, Dec 3, 2008 at 4:46 PM, sas 9 bi user <sas9bi@gmail.com> wrote:

> /*
> Friends,
>
> I have the below data, I need to reshape it and have spent far too long
> trying and thought you all could assist.
>
> The data is survey data, so a person can take multiple surveys during one
> phone call.  Each person is given a unique person ID and then a sequence
> number.  There is also a survey code to differentiate the survey and then
> their responses (string values).
>
> This is what the data looks like in the 'Have' dataset below.  In my 'Want'
> dataset that is how I want it to look when converted to wide.  Some
> questions can have null responses which is reflected in my last record in
> my
> have below.  I need to allow for null responses.
>
> Thanks in advance for your suggestions!
> */
> data have;
>  length type_of_question $10 response $20;
>  input id seq question_number type_of_question $ response $;
>  cards;
>  100 1 1 Health 3
>  100 1 2 Health Good
>  100 1 1 Family 2
>  100 1 2 Family Great
>  100 2 1 Health 1
>  100 2 2 Health Bad
>  100 2 1 Family 9
>  100 2 2 Family
> ;
> run;
> data want;
>  length Health_response_1 $20 Health_response_2 $20 Family_response_1 $20
> Family_response_2 $20;
>  input id seq Health_response_1 Health_response_2 Family_response_1
> Family_response_2;
>  cards;
>  100 1 3 Good 2 Great
>  100 2 1 Bad 9
> ;
> run;
>