data have; input Name $ Date:date9.; format date date9.; datalines; Ben 31Dec2013 Ben 31Jan2014 Ben 31Mar2014 Ben 30Apr2014 Ben 30Jun2014 Ken 31Mar2014 Ken 31May2014 ; run; data want; merge have have(rename=(name=_name date=_date) firstobs=2); if name=_name then do; do i=1 to intck('month',intnx('month',date,0),intnx('month',_date,0)-1); want_date=intnx('month',date,i);output; end; end; format want_date date9.; keep name want_date; run; ... View more

No DO loop necessary, the data step loops over observations on its own. Use a retained variable to keep the start point for each period: data have; input account default; datalines; 12345 2 12345 3 12345 15 12345 16 12345 35 23456 1 ; data want; set have; by account; retain start; if first.account then do; start = default; output; end; else do; if default gt start + 10 then do; output; start = default; end; end; drop start; run; ... View more