^ is not so ^= not equal to   ... View more

the way date is stored in SAS as that it starts from 1/1/1960. So it is not fitting in your calculation. So when you are calculating with date it does not work. below thing may work   data want; set want; numeric=put(ex_post_date,best8.); run; data wantfinal; set want; year=year(numeric); month=month(numeric); day=day(numeric); NEED_num=mdy(month,day,year); format NEED_num mmddyy10.; drop year month day; run; ... View more

I have used the format given by @SASKiwi and used for char variable you have and it worked fine for me data have; informat ex_post_date $9.; input ex_post_date $; cards; 21oct2010 30sep2011 07jan2012 ; run; data want(rename= (ex_post_date1= ex_post_date)); set have; ex_post_date1 = input(ex_post_date,date9.); drop ex_post_date; format ex_post_date1 yymmddn8.; run; ... View more

is your date is currently numeric or in character ... View more

this needs an order by statement. I am not sure what you are looking for. It is missing peoper order which is there in @art297 code   proc sql; select * from a order by marks; quit; ... View more

something like this data want; set have; if prxmatch("m/\s.{3}$/oi",trim(addr_ln_1))> 0 then cntry='CAN'; else cntry='US'; run; ... View more

use  new_var = input(myvar,yymmdd10.);     and do not forget to format as below or whatever you like   format new_var mmddyy10.; ... View more

you may need to give a else statement. If you can do not give else statement, then you will get null if scenarious you have mentioned does not cover all the records. you should probably do else varnum = somevalue ... View more

something like this in SQL by doing union all for each scenario mentioned.  proc sql; create table want as /*scenario 1*/ select id, result from have group by 1 having count(id) =1 union all /*scenario 2*/ select id, min(result) as result from ( select id, result from have group by 1 having count(id) gt 1 and min(result) >= 30)a group by id union all /*scenario 3*/ select id, max(result) as result from ( select id, result from have group by 1 having count(id) gt 1 and min(result) < 30)a group by id order by id; quit; ;   on date column as min or max value   proc sql; create table want as /*scenario 1*/ select id, result, date format=date9. from have group by 1 having count(id) =1 union all /*scenario 2*/ select id, result, min(date) as date from ( select id, result , date from have group by 1 having count(id) gt 1 and min(result) >= 30)a group by id union all /*scenario 3*/ select id,result, max(date) as date from ( select id, result, date from have group by 1 having count(id) gt 1 and min(result) < 30)a group by id order by id; quit; ... View more