Below the response you should see like accept solution, if you click that it is considered as marked. ... View more

Please try   proc sort data=have; by id a; run; data want; set have; by id a; if last.id; run; ... View more

Consider that you have  a period dataset with period variable and start and end date, then you can merge that dataset with main dataset on period then the end date from period could be considered as last period date   Data have ; format cniv1_cniv2 $11. period 2. nbper 8. ; input cniv1_cniv2 period nbper; cards; 10000_20000 4 26 30000_40000 9 34 50000_60000 2 32 70000_80000 1 17 ; run ; data period; input period start :ddmmyy10. end : ddmmyy10.; cards; 1 01/03/2019 31/03/2019 2 01/02/2019 28/02/2019 3 01/01/2019 31/01/2019 ; proc sort data=have; by period; run; proc sort data=period; by period; run; data want; merge have(in=a) period(in=b); by period; if a; format end date9.; run; ... View more

proc sql; create table t1 (id int, typ int, dt char(8), qty int); insert into t1 values (1, 2, '20190325', 50) values (2, 2, '20190320', 15) values (3, 3, '20190401', 50) values (4, 3, '20190405', 5) values (5, 1, '20190406', 25) values (6, 2, '20190411', 5) values (7, 3, '20190412', 15); create table t2 as select id, a.typ, input(dt,yymmdd8.) as date format=date9.,dt, qty from t1(where=(input(dt,yymmdd8.) < '10Apr2019'd )) a inner join ( select typ, max(input(dt,yymmdd8.)) as MaxDate from t1(where=(input(dt,yymmdd8.) < '10Apr2019'd)) group by typ ) b on input(a.dt,yymmdd8.)=b.MaxDate; quit; ... View more