The hash table is the best suited I reckon.
Assuming data set sorted and organised so that ID1 is smaller than ID2.
data HAVE;
input ID1 ID2 ;
cards;
1 3
2 3
5 9
6 9
;
run;
data _null_;
retain GRP;
if _N_=1 then do;
ID=.; GROUP=.;
dcl hash H(ordered:'a');
H.definekey('ID');
H.definedata('ID','GROUP');
H.definedone();
end;
set HAVE end=LASTOBS;;
RC1=H.find(key:ID1);
if RC1=0 then GRP1=GROUP;
RC2=H.find(key:ID2);
if RC2=0 then GRP2=GROUP;
if RC1=0 and RC2=0 then return;
else if RC1=0 then do;
H.add(KEY:ID2, data:ID2, data:GRP1);
end;
else if RC2=0 then do;
H.add(KEY:ID1, data:ID1, data:GRP2);
end;
else do;
GRP+1;
H.add(key:ID1, data:ID1, data:GRP);
H.add(key:ID2, data:ID2, data:GRP);
end;
if LASTOBS then H.output(dataset:'WANT');
run;
proc print data=WANT;
run;
Obs
ID
GROUP
1
1
1
2
2
1
3
3
1
4
5
2
5
6
2
6
9
2
... View more
