concatenate all columns having same starting name

[Nasser_DRMCP]

Hello,

I have a data set that holds 20 columns classe_1, classe_2, classe_3....classe_20.

possibles classe values : 5, 4, 17, C, B2.

i would like to concatenate all columns that begins with class_

thanks for your help in advance

regards

Nasser

[novinosrin]

var=cats(of class_:);

[ballardw · Accepted solution]

From the documentation cats and related concatenation functions will default to a length of 200 characters if not explicitly assigned prior to use. So if your 20 class variables have more than 10 characters each you may need to assign a maximum expected length.

Also if you want a space or comma between the classes then use Catx

var= catx(',', of class: );

or (space between)

var= catx(' ', of class: );

or (and especially if people are to read this result) a comma and a space

var= catx(', ', of class: );

[Nasser_DRMCP]

hello novinosrin

thanks for your help. but I get this message "The CATS function call does not have enough arguments" by attempting your suggestion.

however it works with

var= catx('', of class: );

[Astounding]

That's because your actual variable names are classe_1, classe_2, etc., not class_1, class_2.  This would fix it (but would still have the issue of possibly removing too many blanks):

var = cats(of classe_: );

[emrancaan]

You can use this macro,  just add you DATASET NAME and LIBRARY NAME into it.

%macro getvars(lib,dsn);
    %global vlist;
    proc sql;
        select name into :vlist separated by ' '
        from dictionary.columns
        where memname = upcase("&dsn")
		and libname =upcase("&lib")
		and name like 'class%'
    ;quit;
%mend;

%getvars(LIBRARY,DATASET);

%Put &vlist