Remove duplicate string with condiction?

[ybz12003]

Hello:

I would like to remove the group containing (10, 20, 30, and 40) duplicates with the stings the same in Class, Component and Product.   If the strings in Class, Component and Product are different, I would like to keep it.  Therefore, my final result should be something list below.   Please advice.   Thanks.

data test;

infile datalines dsd;

input Group Class : $20. Component : $20. Product : $20.;

datalines;

10, ANTI/INSULIN, , ,

10, ANTI/INSULIN, , ,

10, ANTI/VITA, , ,

15, PHENAZO, , ,

15, PHENAZO, , ,

19, Large, , ,

20, , HERBS, ,

20, , PSYCH, ,

20, , PSYCH, ,

22, , SKIN, ,

27, , HERBS, ,

27, , HERBS, ,

30, ANTI/INSULIN, , PHENAZO,

30, ANTI/INSULIN, , PHENAZO,

40, , , FLOWER,

40, , , SUBST,

45, , , BANZOCAINE,

45, , , BANZOCAINE,

;

data final;

infile datalines dsd;

input Group Class : $20. Component : $20. Product : $20.;

datalines;

10, ANTI/INSULIN, , ,

10, ANTI/VITA, , ,

15, PHENAZO, , ,

15, PHENAZO, , ,

19, Large, , ,

20, , HERBS, ,

20, , PSYCH, ,

22, , SKIN, ,

27, , HERBS, ,

27, , HERBS, ,

30, ANTI/INSULIN, , PHENAZO,

40, , , FLOWER,

40, , , SUBST,

45, , , BANZOCAINE,

45, , , BANZOCAINE,

;

[RW9]

Is that really your question, as I would think:

proc sort data=test nodupkey;
  by group class;
run;

Would suffice?

[ybz12003]

Hi RW9:

I apologized that I didn't make my sample dataset clearly.   I only need to remove the duplicated in (10, 20, 30, 40).  In other word, if the duplicates are in (15, 27, and 45), I would like to keep them.   Please see my adjusted post above.   Thanks. 

[RW9 · Accepted solution]

Then subset + sort:

data temp;
  set have;
  where group in (10,20,30,40);
run;
proc sort data=temp nodupkey;
  by group class;
run;
data want;
  set have (where=(group not in (10,20,30,40)) temp;
run;

[ybz12003]

Awesome! thanks.