Managing Multiuser Databases Questions and Answers

Answer: D

All five are lock modes, so the answer is (i), (ii), (iii), (iv) and (v).

Multiple granularity locking adds intention modes to the ordinary shared (S) and exclusive (X) locks, so that a transaction can lock a whole file or a single record consistently. IS (intention-shared) and IX (intention-exclusive) on a node announce shared or exclusive locks lower in the tree, and SIX (shared and intention-exclusive) locks a node in shared mode while signalling exclusive locks below.

Each of the narrower options leaves out one or more of these standard modes.

Answer: B

Statements (i) and (ii) are correct, so the answer is (i) & (ii).

(i) A blind write is exactly a write to a data item that the transaction has not read first. (ii) In wound-wait, an older transaction wounds (rolls back) a younger one, but a younger transaction simply waits for an older one; this is generally reckoned to cause fewer rollbacks than wait-die.

(iii) is false: in wait-die it is precisely the younger transaction that dies (is rolled back) when it requests an item held by an older one. (iv) is false: two-phase locking uses shared (read) and exclusive (write) locks, with unlock as an operation, not a lock mode, and 'shared' and 'read' are the same lock. Any option including (iii) or (iv) is therefore wrong.

Answer: A

A transaction becomes partially committed when it finishes the execution of its final statement.

The transaction states are active, partially committed, committed, failed and aborted. After the last statement executes, the transaction is partially committed: its effects may still be only in memory buffers. It becomes committed once enough information has been written to stable storage, which is when the commit record reaches the log.

Starting the first statement puts a transaction in the active state, writing COMMIT to the log marks the committed state, and every successful transaction passes through partial commit.

Answer: B

Correct answer: False

Locks placed by command are explicit; implicit locks are those the DBMS places on its own initiative to enforce the isolation level in force. The two terms classify locks by who asked for them, which is a different question from whether the lock is shared or exclusive.

Answer: B

Correct answer: False

The term is lock granularity, not lock sizing. It describes whether the lock covers a row, a page, a table or the whole database, and the choice trades concurrency against the overhead of tracking a large number of locks.

Answer: B

Correct answer: static.

A static cursor takes a snapshot of the data when it is opened and shows that snapshot throughout, so changes made by others afterwards are invisible. It is the opposite of the dynamic cursor, and the keyset cursor sits between them, showing changes to the rows originally selected but not rows added later.

Answer: B

Correct answer: explicit locks

An explicit lock is one the application asks for by issuing a command; an implicit lock is one the DBMS places on its own initiative to enforce isolation. The distinction is about who requests the lock. Shared and exclusive describe something different - what the lock permits others to do - so the two classifications are independent.

Answer: A

Correct answer: implicit locks

Locks placed by the DBMS on its own initiative, to enforce the isolation level in force, are implicit. An explicit lock is one the application requests by command. Most locking in practice is implicit, which is the point of declaring an isolation level: the developer states what must be prevented and leaves the DBMS to arrange it.

Answer: B

Correct answer: False

Oracle has no Currency datatype. Money is held in a NUMBER column with an explicit precision and scale, such as NUMBER(9,2), which stores the value exactly rather than in binary floating point - and exactness is what matters for money. The Currency type belongs to other products.

Answer: A

Correct answer: True

The four properties are atomic, consistent, isolated and durable. Atomic means all or nothing; consistent means the database moves from one valid state to another; isolated means concurrent transactions do not see each other’s intermediate states; and durable means committed changes survive failure.