finding if an observation happened before a reference date

[lillymaginta]

 I have the following data, multiple observation per id. several s_Date, one adm_date per id,

I want to do the following: if hf=1 and the s_date is before adm_date then delete all observation for that id. 

id       s_date           hf      adm_Date         
1        1/1/2005       1       2/2/2004
1        7/7/2005       0
2        4/3/2004       0
2       2/2/2005        1       6/7/2005
2                       0
3                       0
3       4/4/2006        1     8/8/2005
3

[mohamed_zaki]

Is there will always be only one hf=1 per id?

[lillymaginta]

yes one hf per id

[mohamed_zaki · Accepted solution]

data have;
input id s_date ddmmyy9. hf adm_Date ddmmyy9.;
cards;
1 1/1/2005 1 2/2/2004
1 7/7/2005 0 .
2 4/3/2004 0 .
2 2/2/2005 1 6/7/2005
2 . 0 .
3 . 0 .
3 4/4/2006 1 8/8/2005
;
run;

proc sql ;
create table want as
select *
from have
where id not in (select distinct id from have where s_date < adm_date and hf=1)
;
quit;

[stat_sas]

proc sql;
create table want as
select * from have
group by id
having sum(case when s_date<adm_date and hf=1 then 1 else 0 end)=0;
quit;

[lillymaginta]

Thank you Mohamed asnd stat_sas both codes worked and produced the same results. I appreciate it