topic Re: merging data with only 1 variable each in SAS Programming

merging data with only 1 variable each

stellapersis7 — Sat, 01 Oct 2022 17:41:39 GMT

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

Re: merging data with only 1 variable each

Kurt_Bremser — Sat, 01 Oct 2022 21:26:41 GMT

Just use WHERE:

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

Re: merging data with only 1 variable each

stellapersis7 — Mon, 03 Oct 2022 02:01:37 GMT

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

Re: merging data with only 1 variable each

Tom — Mon, 03 Oct 2022 02:41:09 GMT

@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;

Re: merging data with only 1 variable each

SubbuPaz — Mon, 03 Oct 2022 16:02:18 GMT

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;