Date:         Fri, 24 Sep 2010 12:09:59 -0500
Reply-To:     "Data _null_;" <iebupdte@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Data _null_;" <iebupdte@GMAIL.COM>
Subject:      Re: PROC REPORT: Across - remove cols of all missing values
Comments: To: matt.pettis@thomsonreuters.com
In-Reply-To:  <4698336393F47347A088FB9F99FF1EBACF8999@TLRUSMNEAGMBX26.ERF.THOMSON.COM>
Content-Type: text/plain; charset=ISO-8859-1

PROC statement option NOCOMPLETECOLS

On Fri, Sep 24, 2010 at 12:01 PM, Matthew Pettis
<matt.pettis@thomsonreuters.com> wrote:
> Hi,
>
> I am making a report that is simulated in the code below, and I'd like
> to eliminate columns that have no data.  In the case below, I have the
> following situation:
>
> 1. When level_1 = 'I', then level_2 is either 'A' or 'B'.
> 2. When level_1 = 'II', then level_2 is either 'C' or 'D'.
>
> When I make my report as shown below, I get columns for when level_1 =
> 'I' and level_2 = 'C' or 'D' and for level_1 = 'II' and level_2 = 'A' or
> 'B'.  These don't have data, and have a column of all missing values.
>
> Is there a way to elimitate columns that are all missing that arise in
> situations like this?
>
> Thanks,
> Matt
>
>
>
> === Begin SAS ===
> data have;
>    infile datalines;
>    input level_1:$8. level_2:$8.;
>
>    do i=1 to 3;
>        metric_1 = ranuni(0);
>        metric_2 = ranuni(1);
>        output;
>    end;
> datalines;
> I A
> I B
> II C
> II D
> ;
> run;
>
>
> proc report data=have nowd headline headskip ls=192 ps=330;
>    columns
>        i
>        level_1,level_2,(metric_1 metric_2)
>    ;
>    define i / group;
>    define level_1 / across;
>    define level_2 / across;
>    define metric_1 / analysis format=8.3;
>    define metric_2 / analysis format=8.3;
> run;
> === End SAS ===
>