I'm trying to figure out how to see if a decomposition is dependency preserving or not. The relation is: R(ABCDEF)and have the following FD's. AB -> CE,C -> EB,E -> D,C -> D. We then split the relation into: R1(BF), R2(ACB) and R3(CDE). Is this dependency preserving?
I was under the impression that to calculate this, you do a closure on all of the left sides of the FD's. This gives:
AB+ = ABCEBD which includes AB -> CE
C+ = CEBD which includes the FD's
E+ = ED which include E->D
So in my world, this is dependency preserving. However, according to the markings the answer is that it isn't. What am I doing wrong and/or misunderstanding about the concept?
I do understand that some of the dependencies don't hold in each decomposed relation. For example AB -> E since we can't find a relation which includes these three together. However, I though that since the closure of AB still includes E it would be dependency preserving anyway. Is this where I go wrong? What is an explanation of the concept? (My textbook is VERY brief.)