Date:         Thu, 15 Apr 2004 16:25:13 +0100
Reply-To:     Ari Toikka <toikkari@YAHOO.CO.UK>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ari Toikka <toikkari@YAHOO.CO.UK>
Subject:      Re: cleaning up a dataset
In-Reply-To:  <C93AE4021921784ABAE89216FF4C371803D2CE54@ROC2T7.ghc.org>
Content-Type: text/plain; charset=iso-8859-1

Hi,

actually the solution proposed by Roy Pardee is the simplest:

options missing=' ';
proc transpose data=test out=test2;
   var _all_;
run;

data test2(keep=_name_);
   set test2;
   if col1=' ';
run;

proc transpose data=test2 out=test3(drop=_name_ _label_);
   var _all_;
run;


Ari Toikka
Statistics Finland



"Pardee, Roy" <pardee.r@GHC.ORG> wrote:How about:

- PROC TRANSPOSE your data so you've got one obs per var.
- Filter out the missing values in the resulting dset.
- PROC TRANSPOSE your data back to one rec w/many vars.

If I had more time (or if you'd included code for reading in a sample
dataset in your original post 8^) I'd have coded that up more
concretely...

HTH,

-Roy

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
William Kossack
Sent: Wednesday, April 14, 2004 3:30 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: cleaning up a dataset


except that I need it in a dataset.

my current code looks something like this &&attrib&x is a macro variable
that lists all character fields (I've split off the character data for
simplicity)
I can create a dataset with only character data but when I go to
reference &&attrib&x later I run into problems

data characterdata (keep= %do x=1 %to &amax; &&attrib&x %end;); set f;

%do x=1 %to &amax;
%let value = &&attrib&x;
%if &value ^= "" %then %do;
drop &&attrib&x;
%end;
%end;


"Copeland, Laurel" wrote:

> only 1 record? this will write to the log...
> DATA missingjunk; SET yourdataset;
> ARRAY NUM _NUMERIC_; DO OVER NUM; IF NUM=. THEN DO; PUT NUM=; END;
> END; ARRAY CH _CHARACTER_; DO OVER CH; IF CH=' ' THEN DO; PUT CH=;
> END; END; RUN;
>
> if more than 1 record, include the id var in the PUT
> PUT idvar= num=;
>
> hth, laurel
>
> -----Original Message-----
> From: William Kossack [mailto:kossackw@NJC.ORG]
> Sent: Wednesday, April 14, 2004 5:12 PM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: cleaning up a dataset
>
> I have a dataset with only one record both character and numberic
> fields.
>
> I'm trying to find a way of keeping only the variables that missing
> data so a report can be generated showing where information is missing


---------------------------------
  Yahoo! Messenger - Communicate instantly..."Ping" your friends today! Download Messenger Now