Date:         Tue, 24 Dec 1996 08:09:04 -0800
Reply-To:     "Derrick J [Rick] Bates" <DJ_Bates@PNL.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Derrick J [Rick] Bates" <DJ_Bates@PNL.GOV>
Organization: Battelle, Pacific-Northwest National Laboratory,  Richland, WA,
              USA
Subject:      Re: Problem with converting decimal numbers to character text

In article <32bf6e7e.79584@nntp.ix.netcom.com>, BentleyJ@ix.netcom.com
(John E. Bentley) wrote:
 
> I have numeric variables like 96.1 and 95.7 that I need to convert to
> character variables so that I can recode missing to 'na'.
>
> I've specified the length of the new character variable as $8.
> When I use
>
>   charvar=put(numvar,$8.)
>
> it rounds the numeric variable.  96.1 is output as 96 and 95.7 is
> output as 96.  How can I keep this from occurring?
>
> Can anyone suggest code that will do what I want done?
>
 
   What is it you want to do
 
   If all your non-missing are of the form you give above you might use
      length charvar $8; charvar=numvar;   * automagic formatting;
         or
      length charvar $8; charvar=put(numvar,8.1); * specified formatting;
         or
      etc., etc., etc.
 
   Another option for getting the 'na' without going to character variables
   is using PROC FORMAT;
      proc format;
      value na . = '      na';
      run;
   then use
      format numvar na.;
   when printing or outputting to a file for S to read
 
-----------------------------------------------------------------
 Derrick J. (Rick) Bates              | mailto:DJ_Bates@PNL.Gov
 Environmental Statistics Group       |  Voice: 509.375.2740
 2400 Stevens Drive, K5-12            |    FAX: 509.375.2604
 Battelle, Pacific-NW National Lab    |
 P.O. Box 999, Richland, WA 99352 USA | EVERY POSSIBLE DISCLAIMER
------------------------------------------------------------------