Dropping Unnecessary Variables When Building a SAS Data Set from Raw Data

Which is more efficient?

Code 1:

Data Smaller;
	Set BigOne;
	Keep X1-X10 NewMean;
	NewMean=Mean(of X1-X10);
run;

Code 2:

Data Smaller;
	Set BigOne (Keep=X1-X10);
	NewMean=Mean(of X1-X10);
run;