Date:         Thu, 24 Jul 2008 23:40:35 -0400
Reply-To:     "Howard Schreier <hs AT dc-sug DOT org>"
              <schreier.junk.mail@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Howard Schreier <hs AT dc-sug DOT org>"
              <schreier.junk.mail@GMAIL.COM>
Subject:      Re: create variable names from values of other variable

On Thu, 24 Jul 2008 23:25:34 -0400, Howard Schreier <hs AT dc-sug DOT org>
<schreier.junk.mail@GMAIL.COM> wrote:

>On Thu, 24 Jul 2008 17:31:46 -0700, learner <cnfengshuang@GMAIL.COM> wrote:
>
>>Hi,
>>
>>What I want to do is like this. Suppose I have some data as below.
>>
>>id      date    value
>>1       200801  100
>>2       200802  101
>>3       200803  99
>>
>>I want to create a new variable each time value>=100, with part of the
>>variable name record the date. like this:
>>
>>id      date    value   value_200801    value_200802
>>1       200801  100     100             101
>>2       200802  101     100             101
>>3       200803  99      100             101
>>
>>I have a large data set to process like this, so please help me to
>>find an easy way to do it.  I am going to do numerical analysis for
>>each id, and need to drop the added variables in the process of
>>calculation before processing the next id,  to save system resource.
>>Can I do this in the DATA step? or is IML better?
>>
>>TIA
>>Learner
>
>Try this:
>
>   data have;
>   informat id $1. date yymmn6.;
>     format        date yymm7.;
>   input
>id      date    value  ; cards;
>1       200801  100
>2       200802  101
>3       200803  99
>   ;
>
>   data ge100;
>   set have;
>   where value ge 100;
>   run;
>
>   proc transpose data=ge100 out=wide(drop=_name_) prefix=value_;
>   id date;
>   var value;
>   run;

Or combine the above 2 steps as:

   proc transpose data=have out=wide(drop=_name_) prefix=value_;
   where value ge 100;
   id date;
   var value;
   run;

>
>   data want;
>   set have;
>   if _n_=1 then set wide;
>   run;