Using them as column headers makes no sense, as you can't change your dataset structure halfway through the data step.
Instead you define the columns with the input statement, and you best simply discard those lines:
data test;
infile in truncover;
input x1 $ x2 $;
if upcase(substr(x1,1,1)) ne 'X';
run;
If you only had to read one file, using firstobs=2 in the infile statement will also do the trick.
... View more
