Keep variables with different values among observations

[Mathis1]

Hi, 
I have this tiny table attached (3 observations) with a lot of variables. I'd like to keep only the variables for which there exists different values between at least 2 observations (or supress variables that have the same values for all the observations). Do you know a simple way to do this ? 

[Tom]

The NLEVELS output of PROC FREQ will show you how many distinct values each variable has.

You can use ODS OUTPUT to get the result into a dataset.

ods output nlevels=nlevels;
proc freq data=carross1 nlevel;
  tables _all_ / noprint;
run;

You can now pull the list of names that mean you criteria into a macro variable using PROC SQL.  Either make a list of the ones to keep or the ones to delete and use the list in a KEEP or DROP statement (or dataset option).

[smantha]

proc sort data=inputdata nodupkey;
by _all_;
run;