Separating out observations for mismatch

[r3570]

I am having 2 datasets as below

dataset 1:

SUBJID	VISIT
1002	Baseline
1002	DAY 1
1002	DAY 2
1002	DAY 3
1003	Baseline
1003	DAY 1
1003	DAY 2
1003	DAY 3
1004	Baseline
1004	DAY 1
1004	DAY 2
1004	DAY 3

dataset 2:

SUBJID	VISIT
1002	Baseline
1002	DAY 1
1002	DAY 2
1003	Baseline
1003	DAY 1
1003	DAY 2
1003	DAY 3
1004	DAY 1
1004	DAY 2
1004	DAY 3

I want to separate out the records from dataset 1 for which there is no corresponding VISIT in dataset 2.

Need help with the code. Can anybody help ?

[Kurt_Bremser · Accepted solution]

Since the datasets are already nicely sorted, a MERGE will do:

data want;
merge
  dataset1
  dataset2 (in=ds2)
;
by subjid visit;
if not ds2;
run;