join or merge by using some information

[Lily07]

data have1;
 input id year death reg;
 datalines;
1 2000 0 11
1 2001 0 12
1 2002 0 11
1 2003 0 11
2 2000 1 11
2 2001 1 12
2 2002 1 13
2 2003 1 14
3 2000 0 . 
3 2001 0 11
3 2002 0 11
3 2003 0 11
5 2000 1 . 
5 2001 1 . 
5 2002 1 11
5 2003 1 11
;

data have2;
input id year reg sex incs;
datalines;
1 2000 11 1 1000
2 2000 11 1 2000
4 2000 12 1 4000
;

data want;
input id year reg death sex incs;
datalines;
1 2000 11 0 1 1000
1 2001 12 0 1 1000
1 2002 11 0 1 1000
1 2003 11 0 1 1000
2 2000 11 1 1 2000
2 2001 12 1 1 2000
2 2002 13 1 1 2000
2 2003 14 1 1 2000
4 2000 12 . 1 4000
;

I'd like to have a dataset like want.

The logic is that I want to keep IDs only from have2, but want to keep all person-year information for the IDs in have1. Except for the reg variable, sex and incs are time-invariant information, so I want to have the same values for all years, but in case of reg it's all changing over years, so I want to keep information from have1.  Meanwhile, there may be some IDs only shown in have2, in that case, I want to keep only the first row (as shown in have2) not making other years.

Thank you so much in advance.

[CarmineVerrell]

This code seems to give me the want dataset . Take a look.


Data thewant;
merge have1(in=in1)
have2(in=in2);
by id;
if in2;
run;

[Shmuel]

According to ID=4 wanted output it seems that you select DEATH only from have1 and neglect it from have2. Try next code:

proc sql;
  create table want as
  select a.ID, a.year, a.reg, a. sex, a.incs,
            b.death
  from have2 as a
  left join have1 as b
  on a.ID = b.ID;
quit;