You are correct that 'x=" "' will work for numeric variables, but SAS provides a function, MISSING, that tests a variable or expression (numeric or char) and returns a 1 if the value of the var is missing and 0 if not. Using the function will prevent the internal conversions and note generations.

----- Original message ----- From: "Ari Toikka" <toikkari@YAHOO.CO.UK> To: SAS-L@LISTSERV.UGA.EDU Date: Thu, 15 Apr 2004 16:04:52 +0100 Subject: Re: cleaning up a dataset

Hi,

/* take the variable names */ proc contents data = test out= cont(keep=name); run;

/* generate macro variable keeppgm, which holds program */ /* statements which generate a keep list */ /* if x=" " is valid for numeric variables too */ data _null_; length keeppgm $ 2000; set cont nobs=records; keeppgm = substr(keeppgm,1,length(keeppgm)) !! 'if ' !! compress(name) !! '=" " then keeplist = substr(keeplist,1,length(keeplist)) !! " " !! "' !! compress(name) !! '";'; retain keeppgm; if _n_ = records then call symput('keeppgm', keeppgm); run;

/* generate the keep list and use it */ data test2; length keeplist $ 2000; set test; &keeppgm; call symput('keeplist',keeplist); keep &keeplist; run;

Ari Toikka Statistics Finland

William Kossack <kossackw@NJC.ORG> wrote: 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

-- http://www.fastmail.fm - The way an email service should be