libname hercules 's:\courses\stat-renaes\stat404';

proc print data=hercules.nonsales;
run;

proc print data=hercules.nonsales;
   var Employee_ID Gender Salary Job_Title
       	Country Birth_Date Hire_Date;
   where Employee_ID = . or
        Gender not in ('F','M') or
        Salary not between 24000 and 500000 or
        Job_Title = ' ' or
        Country not in ('AU','US') or
        Birth_Date > Hire_Date or
        Hire_Date < '01JAN1974'd;
run;

* you could also use a DATA step to create a new dataset that meets the 
conditions set in the WHERE statement of PROC PRINT (the first methods requires
2 less lines of code -- but same results.  Although you may not want to create a dataset
for validation.;

data work.emps;
    set hercules.nonsales;
    var Employee_ID Gender Salary Job_Title
       	Country Birth_Date Hire_Date;
    where Employee_ID = . or
        Gender not in ('F','M') or
        Salary not between 24000 and 500000 or
        Job_Title = ' ' or
        Country not in ('AU','US') or
        Birth_Date > Hire_Date or
        Hire_Date < '01JAN1974'd;
run;

proc print data=emps;
run;

proc means data=hercules.nonsales;
   var Salary;
run;

proc means data=hercules.nonsales n nmiss min max;
   var Salary;
run;

proc univariate data=hercules.nonsales;
   var Salary;
run;

proc freq data=hercules.nonsales;
   tables Gender Country;
run;

proc freq data=hercules.nonsales;
   tables Employee_ID;
run;

proc freq data=hercules.nonsales nlevels;
   tables Gender Country Employee_ID;
run;

proc freq data=hercules.nonsales nlevels;
   tables Gender Country Employee_ID / noprint;
run;

proc freq data=hercules.nonsales nlevels;
   tables _all_ / noprint;
run;