Date:         Mon, 1 May 2000 23:23:11 GMT
Reply-To:     sashole@mediaone.net
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Paul Dorfman <paul_dorfman@HOTMAIL.COM>
Subject:      Re: Reshaping data?
Comments: To: bja-bob@MINDSPRING.COM
Content-Type: text/plain; format=flowed

Bradley,

Judging from your output, you just have to repeat the record 30 times:

data vert (keep=id day var:);
   set hor;
   do day=1 to 30;
      output;
   end;
run;

Note that there is a slight inconsistency between your input as shown and
the desired output: There is no dependency on the way the days are valued.
No matter whether the boolean variable day is set to 1 or 0, the record
winds up in the output. In the case you might want to output only records
corresponding to the non-zero days, the code above should be somewhat
amended:

data vert (keep=id day var:);
   array d(*) day1-day30;
   set hor;
   do day=1 to 30;
      if d(day) then output;
   end;
run;

Kind regards,
=====================
Paul M. Dorfman
Jacksonville, Fl
=====================

>From: "Bradley J. Anderson" <bja-bob@MINDSPRING.COM>
>Reply-To: "Bradley J. Anderson" <bja-bob@MINDSPRING.COM>
>To: SAS-L@LISTSERV.UGA.EDU
>Subject: Reshaping data?
>Date: Mon, 1 May 2000 18:26:02 -0500
>
>I need to "reshape" a SAS data set and am having problems figuring out how
>to do this efficiently.  Assume the original data are as follows:
>
>ID  day1 day2 day3 ..... day 30, var1 var2 var3 ... var30
>1      1     0        1               1          10    12    9          23
>  2     1     1        0               1            9     11    7
>21
>3      0     0        1               0           11     10   9         22
>.
>.
>.
>N
>
>I want to reshape and expand the data so that there will be 30 observations
>for each of the cases in the original data set.  The reshaped data set will
>have N*30 observations.  Such as:
>
>ID  Day  Var1  Var2 ...
>1      1       10     12
>1      2       10     12
>1      3       10     12
>.        .
>1     30      10     12
>.
>2      1          9     11
>2      2          9     11
>.        .           .        .
>2     30         9     11
>.
>N      1 .....
>N      2 .....
>
>In short, I want to create 30 days worth of observations (days will become
>the unit of analysis) for each of the orginal cases.  A colleague using
>Stata has been using a combination of a "reshape" command and an "expand"
>command to reshape such data appropriately.  I'm trying to convert what my
>colleague has done into a SAS program.  It's has to be pretty easy, but I'm
>stumbling on this one.  I'd be most appreciative if anyone can offer an
>efficient solution.  Thanks in advance.

________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com