Merge Problems

[Tom]

So if B has 581 observations and 34 do not match any ids from A that leaves 547 that do match. The count of 551 matched records means that four extra observations were added by merging with A.  So either one id in B matched to 5 observations in A or some other combination.  To find which ids are appearing in multiple times you can use the FIRST. and LAST. variables that the BY statement will generate.

data check;
  merge alldata (in=in1) pubdata3 (in=in2);
  by course;
  if in1 and in2 and not (first.course and last.course);
run;
proc print;
run;

[piyushas]

OMG thank you! Yup that's it.  I introduced duplicates further down in my code, after doing my check for duplicates, and did not catch it.

I've been mucking with this for days -thank you!