Different types of merging using Data Step or Proc SQL in

SAS
mehtahemal.com /sas/different-types-of-merging-using-data-step-or-proc-sql-in-sas/450/

By Hemal
Mehta

This post shows different type of merging using Data step and Proc SQL in SAS. Use the following two datasets (A &
B) and give it a try!!

Dataset A = Red color;

Dataset B = Green color;
Common observations in A and B = Yellow color (Red + Green = Yellow!)

DATA A DATA B

Data A; Data B;
input ID Gender$; input ID Sex$;
Datalines; Datalines;
1M 6F
2M 7F
3M 8M
4M 9M
5M 10 M
6F 11 F
7F 12 F
8M 13 F
9M 14 F
10 M 15 F
; ;
run; run;

Graphical
What do you want? representation Data Step Proc SQL

All observations data M; Proc sql;

merge A (in = x) B (in = y); create table M2 as
by id; select coalesce(a.id, b.id) as id,
if x = 1 or y = 1; gender, sex
run; from A full join B
on a.id = b.id;
quit;

All matching obs and non- data N; Proc sql;

matching from A merge A (in = x) B (in = y); create table N2 as
by id; select coalesce(a.id, b.id) as id,
if x = 1; gender, sex
run; from A left join B
on a.id = b.id;
quit;

1/3
 Graphical
What do you want? representation Data Step Proc SQL

All matching obs and non- data O; Proc sql;

matching from B merge A (in = x) B (in = y); create table O2 as
by id; select coalesce(a.id, b.id) as id,
if y = 1; gender, sex
run; from A right join B
on a.id = b.id;
quit;

Only matching observations data P; Proc sql;

merge A (in = x) B (in = y); create table P2 as
by id; select coalesce(a.id, b.id) as id,
if x = 1 and y = 1; gender, sex
run; from A inner join B
on a.id = b.id;
quit;

Only nonmatching from A data Q; Proc sql;

merge A (in = x) B (in = y); create table Q2 as
by id; select coalesce(a.id, b.id) as id,
if x = 1 and y = 0; gender, sex
run; from A full join B
on a.id = b.id
where b.id is null;
quit;

More efficient code:

Proc Sql;
create table Q2 as
select coalesce(a.id, b.id) as id,
gender, sex
from A left join B
on a.id = b.id
where b.id is null;
quit;

2/3
 Graphical
What do you want? representation Data Step Proc SQL

Only nonmatching from B data R; Proc sql;

merge A (in = x) B (in = y); create table R2 as
by id; select coalesce(a.id, b.id) as id,
if x = 0 and y = 1; gender, sex
run; from A full join B
on a.id = b.id
where a.id is null;
quit;

More efficient code:

Proc Sql;
create table R2 as
select coalesce(a.id, b.id) as id,
gender, sex
from A right join B
on a.id = b.id
where a.id is null;
quit;

Only nonmatching from A and data S; Proc sql;

B merge A (in = x) B (in = y); create table S2 as
by id; select coalesce(a.id, b.id) as id,
if (x = 1 and y = 0) or (x=0 gender, sex
and y = 1); from A full join B
run; on a.id = b.id
where a.id is null or b.id is null;
quit;

Showing 1 to 7 of 7 entries
© 2017 Default copyright text

3/3