建议，先把数据排序，然后新建一个变量y2，把y2赋值为各ID的第一个值，然后下一个obs y2增1。再做比较year vs. y2，相同者赋值0给新变量flag，否则flag为1。再做一个proc univariate，对flag求和，不等于0者保留。

求解答

data test;
input id year;
cards;
   1    2001
   1    2002
   1    2004
   1    2005
   2    2002
   2    2003
   2    2004
   3    2002
   3    2005
   3    2007
   ;
run;

proc sort data=test; by id year; run;

data test01;
set test;
by id year;
preyear=lag(year);
if not first.id then do;
dif=year-preyear;
if dif ne 1 then output;
end;
run;

proc sql;
create table test02 as
select *
from test
where cats(put(id,best.), put(year, best.)) in (select cats(put(id,best.), put(year, best.)) from test01)
or   cats(put(id,best.), put(year, best.)) in (select cats(put(id,best.), put(preyear, best.)) from test01);
quit;

data test;
input id year;
cards;
   1    2001
   1    2002
   1    2004
   1    2005
   2    2002
   2    2003
   2    2004
   3    2002
   3    2005
   3    2007
   ;
run;

proc sort data=test; by id year; run;

data test01;
set test;
by id year;
preyear=lag(year);
if not first.id then do;
dif=year-preyear;
if dif ne 1 then output;
end;
run;

proc sql;
create table test02 as
select *
from test
where cats(put(id,best.), put(year, best.)) in (select cats(put(id,best.), put(year, best.)) from test01)
or   cats(put(id,best.), put(year, best.)) in (select cats(put(id,best.), put(preyear, best.)) from test01);
quit;