I was wondering if I could tell proc expand not to delete rows in a data set for which there exists only a single observation per by-group. For example
proc expand data=IN out=OUT method=none;
by bgrp1 bygrp2 bygrp3;
will delete all lines in the OUT data set for which there exist only one observation per by-group in the IN dataset. I would prefer that the lines would still be there with the automatic variable TIME having a value of zero (I know that I could use a data step to create the automatic variable TIME).