Date:         Thu, 26 Jan 2006 23:34:51 -0500
Reply-To:     "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Subject:      Re: Formating in the proc means results

On Thu, 26 Jan 2006 13:32:06 -0800, ulagappan@GMAIL.COM wrote:

>Is there a simple way to format the proc means output
>
>proc means data = sashelp.air;

Strange, I don't have SASHELP.AIR in my install (9.1.3 SP3). So I'll use

   data air;
   do d = 0 to today(); e = -d; output; end;
   format d e date9.;
   run;

>run;
>
>
>I get the  statistical results of date variable in SAS date format.
>I have the round about ways to do it.
>Output the means with format.
>However there is a problem. In my case I have 20 variables.
>So the out table is quite long.
>When I transpose, the format is agian lost.
>So I am doing one variable at a time, instead of all 20.
>
>Any help will be appreciated.
>ulag

TABULATE gives you a lot more control. Consider:

   options nocenter FORMCHAR="|----|+|---+=|-/\<>*";

   proc tabulate data=air;
   var d e;
   table f=date9.*(mean std*f=4. min max),
        d e;
   run;

I've deliberately crossed two format specs, since the standard deviation
should not get a date format. A log warning results, but I get what I want:

--------------------------------------------
|                      |    d    |    e    |
|----------------------+---------+---------|
|Mean                  |13JAN1983|18DEC1936|
|----------------------+---------+---------|
|Std                   |     4858|     4858|
|----------------------+---------+---------|
|Min                   |01JAN1960|06DEC1913|
|----------------------+---------+---------|
|Max                   |26JAN2006|01JAN1960|
--------------------------------------------

To avoid the log warning, I could get verbose and hang a format on each of
the 4 stat keywords.