merging data with only 1 variable each

[stellapersis7]

I have a dataset like this. how do i pick just few items like only  V3000 and 225 and corresponding L_O_S and put it in different table.

N obs	A_D	L_O_S
1	V3000	23
2	225	4
3	v3001	65
4	7856	4
5	v240	67

[Kurt_Bremser · Accepted solution]

Just use WHERE:

data want;
set have;
where a_d in ("V3000","225");
run;

[stellapersis7]

Thanks a lot. I have a this kind of dataset where I need to average the L_O_S variables and pick V3000 and 225 against the corresponding average L_O_S? can anyone please help with it?

N obs	A_D	L_O_S
1	V3000	23
2	225	4
3	v3001	65
4	7856	4
5	v240	67

[Tom]

---

@stellapersis7 wrote:

Thanks a lot. I have a this kind of dataset where I need to average the L_O_S variables and pick V3000 and 225 against the corresponding average L_O_S? can anyone please help with it?

N obs	A_D	L_O_S
1	V3000	23
2	225	4
3	v3001	65
4	7856	4
5	v240	67

---

Your description is not very precise.  What does "pick" mean?  What does "against" mean?

What is the answer you want for the data you showed?  How did you get to that answer?

If you want to find averages use PROC MEANS

proc means data=have;
  class a_d ;
  var l_o_s;
run;

or perhaps the MEAN() aggregate function in PROC SQL.

proc sql;
create table want as
  select a_d,mean(l_o_s) as mean_los 
  from have
  group by a_d
;
quit;

[SubbuPaz]

Try this code if you would like to 'pick' a subset of a_d values and get mean_los 'against' each n_obs | a_d:

Note that in your sample data the 'a_d' column is unique. May be your actual data set has multiple observations for each 'a_d' - for average computation. 

data have;
input n_obs a_d $ l_o_s;
datalines;
1 V3000 23
2 225 4
3 v3001 65
4 7856 4
5 v240 67
;

proc sql;
create table want as
select a.n_obs, a.a_d, b.mean_los
from 
(select n_obs, a_d, l_o_s from have
where  a_d in ("V3000","225"))a
left join
(select a_d,mean(l_o_s) as mean_los 
  from have
  group by a_d)b
on a.a_d = b.a_d
;
quit;