/*** TIP00399 ***/
Proc Sort nodup with many variables
I have a dataset with 200+ variables. If I run a proc sort step on it 
with the nodup option, can I be certain that the records that are 
deleted are definitely identical (in all variables) to ones in the 
dataset that remains? (ie does SAS ensure that all variables have the 
same value before the record is deleted?) 

e.g.
proc sort data=mydata nodup; by ID; run;
* where mydata is a dataset with 200+ variables;


The answer is "no". Instead, use

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