The JOINs in general when you use any ON predicate combine ALL of the join records from the first set with EACH of the matches in the other set. So you get this multiplicative behavior. 1 Key in set a and 2 keys in set b = 2 matches.
It may be possible to filter the results but there is no guarantee.
It is always a good idea to provide at least the code you are attempting and better to include example data along with the desired result.
@sgleeson2 wrote:
I have 2 datasets, a and b which I am merging using a left join. Table a has duplicate join keys, but b does not. When I complete the join I am ending up with more records in a than before. I know the duplicate keys are a problem but they are not something I can change. However, I don't understand WHY they are causing the issue.
My understanding is that in cases of duplicate keys in the left table (a), the corresponding key in the right table (b) will simply match to both occurrences in a, without creating an extra records?
Joins are record by record constructs so you are misunderstanding the basic behavior of a join.
