1.

 Dataset

Answer»

Macro is way of automating in sas programs. There are MANY ways to create macro. 

%macro get_var_dist (data_nm , categ_cut_off); /* Taking list of variables in the dataset along with type*/ proc contents data=&data_nm. out=_content_; run; /* NOTE - how data_nm is getting resolved. */ /* Fiding number of variables to define loop */ proc sql noprint; select count(*) into: nvars from _content_; quit; /* Note this is another way of DEFINING macro variables */ %put "No of Variables is  " &nvars. ; /* Note this can be a useful way to detect the value of macro variables.*/ %do i=1 %to &nvars. ; data _null_; set _content_; if _n_ = &i. then do;        call symput("variable", name);        call symput("var_type",type); end; run; /* Note -   1: One more way of defining variables.   Please note here information is being passed from general SAS dataset   to SAS macros using call symput   2: Please see how to run do a loop */ %put "Variable is  " &variable. ; %put "Variable Type is  " &var_type. ; proc sql noprint;        select count(distinct &variable) into: distinct_categories        from &data_nm.;        quit; %put "Number of distinct categories is  &distinct_categories." ; /* Note the syntax of writing if else condition */ /* %do ....  %end; defines the block */ /* Missing command below includes missing value as a category */ /* Also note, if you don't use %eval it can throw errors */ /* if comparison involves floating numbers, you will need to use %sasevalf */ %if %eval(&distinct_categories.) <=  %eval(&categ_cut_off.) %then %do; Title "Freq Dist for Variable &variable. as number of distinct category is &distinct_categories.";   proc freq data=&data_nm. ; table &variable. / missing ; run; %end; %if %eval(&var_type.) =  1 %then %do; Title "Univariate Dist for Variable = &variable. ";   proc univariate data=&data_nm. PLOT; var &variable.  ; run; %end; %end; %mend get_var_dist; options symbolgen mlogic mprint; %get_var_dist(data_nm=sashelp.heart, categ_cut_off=25); %get_var_dist(data_nm=sashelp.asscmgr, categ_cut_off=10);


Discussion

No Comment Found

Related InterviewSolutions