Dependence 1. from primary key to super key a. Primary key b. Candidate key c. Super key 2. Functional dependence a. Domain: a set of attributes b. Codomain: a set of attributes c. Terms: determinant vs. dependent d. Interpretation: XY For each instant of a relation, two tuples agrees on X values implu agree on Y e. Trivial: Y is a subset of X ( projection) f. Note: if X is not a candidate key, we have redundancy on Y. 3. S+=Closure of S a. Use composition b. If it is not easy to find S+ c. Work on a lesser problem w/o having S+ 4. Find all attributes that are dependent on Z ( a set of attributes.) i. Inference rules ( see later) ii. Algorithm iii. Ex: find all attributes of {A,B} that are dependent on: ABC; ECF; BE; CDEF Ans: {ABCDF} : call it AB+ iv. Benefits: 1. superkey 2. given XY how do we know it belongs to S+ w/o knowing S+ (check if Y is a subset of X+ 5. Inference rules a. Reflexive b. Augmentation c. Transitivity d. Self determination e. Decomposition f. Union g. Composition h. General Unification Theorem: If AB and CD then A union (C-B) BD 6. Irreducible sets a. Cover: S2 cover S1 if S1* is a subset of S2* b. S1 and S2 cover each other: they are equivalent. c. Given S, we might have many of its equivalents. Which one we want to select as a representative? Ans: the one with the following properties: i. Right hand of each FD must be ONE attribute ii. Left hand side can not be reduced w/o changing S* iii. No FD’s can be discarded w/o changing S* d. Given an S, can we find its representative? i. Yes ii. Its representaive is called IRREDUCIBLE.