After seeing Andre's response, there is a better understanding of the problem. I don't know enough about using the UNICODE function but it seems to convert Unicode characters to the numeric encoding in the current SAS session. There must be an additional step that I am missing with that solution.

Further clarification of the problem will help. You will need to define what is acceptable as values.

1. Numbers 2. Alphabets 3. You may want to visit http://www.asciitable.com/ and see the range of acceptable values in your variables. You may then try something like the following where __n1 corresponds to a flag for var1, __n2 for var2 and so on. A positive value indicates the presence of non-acceptable characters:

** define a space, a left and a right paren as compress characters ** ; ** adjust accordingly - using asciitable.com ** ; %let compchar=( ) ;

data want ; set have end = eof ; array __c (*) $ _character_ ; array __n (3) ; do _n_ = 1 to dim(__c) ; __n(_n_) = notalnum(compress(strip(__c(_n_)),"&compchar")) ; end ; run ;

options nocenter ; proc print ; run ;

Venky Chakravarthy

On Thu, Nov 3, 2011 at 5:08 PM, Venky Chakravarthy <sasyellvenky@gmail.com> wrote: > Have you tried using the ANYCNTRL function. > > http://support.sas.com/documentation/cdl/en/lefunctionsref/63354/HTML/default/viewer.htm#p1mt339n2zhb33n1uejmw7470orv.htm > > Venky Chakravarthy > > On Thu, Nov 3, 2011 at 4:41 PM, Singh <adsingh78@gmail.com> wrote: >> Hi, >> >> Is there a way to output only those observations and only those variables >> in a SAS dataset that have special characters? >> >> HAVE: >> >> id(N) � � var1(C) � � � � � � var2(C) � � � � � � � � � � var3(C) >> 101 � Klean-Prepā„¢ � � � diabetes � � � � � � � Nutritional Supplement >> 102 � lanzor � � � � � � � Health Suppelment � � � lidocaĆÆne >> 103 � L-Tyrox 50 Āµg � � � fish oil supplement � � lopĆ©ramide (Imodium) >> >> WANT: >> >> id(N) � � var1(C) � � � � � � var3(C) >> 101 � Klean-Prepā„¢ � � � Nutritional Supplement >> 102 � lanzor � � � � � � � lidocaĆÆne >> 103 � L-Tyrox 50 Āµg � � � lopĆ©ramide (Imodium) >> >> I am using the following code, but this does not work as I get error >> message: >> ERROR: Some character data was lost during transcoding in the dataset >> >> data want; >> � set have; >> � if compress(lowcase(var1),'abcdefghijklmnopqrstuvwxyz') eq '' then >> delete; >> run; >> >> >> >> Thanks in advance!! >> >