More / Edge Cases

Is A a candidate key in the earlier 6‑attribute schema?

B. No

Candidate keys:

A. AG

BCNF violation in in_dep:

A. dept_name→building,budget

Canonical cover must (multiple select):

A. Have no extraneous attributes
B. Have unique LHS per FD
C. Preserve closure of original F

3NF synthesis: if no schema contains a candidate key, you must:

B. Add a relation containing a candidate key

Lossy join yields:

C. A superset of the original relation (spurious tuples)

Dependency‑preservation test properties (multiple select):

A. Polynomial time
C. Per FD in F
D. Iterative attribute propagation across fragments

NOT a purpose of normalization:

B. Guarantee best physical performance

Airline weekdays to 1NF:

B. One tuple per (Flight,Weekday)

BorrowedBooks(BookNo,Patron,Address,Due) truths (multiple select):

A. Address depends on Patron
B. Not in 3NF
C. Decompose to separate Patron→Address

Parsing roll numbers (“CS0012” → dept) violates:

A. 1NF atomicity in use

Modeling a set‑valued attribute (multiple select):

A. Separate relation keyed by owner PK
D. Weak entity if it has its own attributes

Lossless for (A,B) and (A,C):

A. A→B or A→C

In 5NF, JD {A,B,C} is implied by candidate keys iff:

A. A, B, C are each superkeys

Testing extraneous on RHS (multiple select):

A. Replace α→β with α→(β−{B})
B. If B is still implied, then B is extraneous
C. Otherwise keep B
D. Test using the canonical cover set

BCNF yet anomalous example:

A. inst_info(ID, child_name, phone)

Decomposition pitfalls (multiple select):

A. Not dependency‑preserving → expensive checks
B. Lossy decomposition → spurious tuples
C. Over‑decomposition can hurt performance
D. 3NF synthesis avoids lossiness and non‑preservation

Putting date on student instead of on advisor causes:

A. Loss of information granularity

Union schema compatibility requires:

B. Same domains by position and same arity

Set operations require (multiple select):

A. Same arity
B. Domain compatibility by position
D. Set (duplicate‑eliminating) semantics