I have read the info on finding duplicates with a self join and numbering them via Count() function, but my situation is slightly different I think.

I have a Table of students who attend my school and I often email their Parents.

I get my found set usually by go to related record from a class they are in or a course they are taking.

When two or more siblings are in my found set, the parent gets the same email for each of them, which is not a major problem, but I'd prefer to send only one email per family.

Eliminating the second family member from the found set before sending doesn't work because sometimes only the second family member (by the self join) is in the found set.

Can anyone suggest a good way to do this?

Thanks & happy holidays to all.

Do you have a table of Families?

No. Only a self join for students called family members based on same home phone and surname.

Do you think I need a families table?

Well, it would help - because by GTRR [Families] you would get the set you're looking for. As it is, I don't see that you have much of a choice, except sort the found records by family and omit all but one record from each group. This could be made a bit simpler (and faster) by using a technique called 'Fast Summaries':

http://www.onegasoft.com/tools/fastsummaries/index.shtml

Another option would be to use an extra relationship and some calculations to de-dupe the set - but since it's going to be scripted anyway, I don't see why you should complicate your schema just for this.

Thanks again Comment. It looks like a Families table is the only way to go.

Phil