OCP 23ai New Features Practice Test 3 — 15 Questions

Question 1

What is the primary benefit of Lock-Free Reservations in high-concurrency environments?

Accepted Answer

Reduced row-level locking contention. It eliminates wait events associated with row-level locking for numeric updates.

Answer

Faster queries

Answer

Better compression

Answer

Automatic indexing

Question 2

Which TWO of the following statements about True Cache are correct?

Accepted Answer

It resides in the application tier. Option D is correct because Oracle True Cache is deployed as a separate, lightweight instance that runs in the application tier, close to the application servers, so it can serve queries with minimal network latency. Option E is correct because True Cache is fundamentally a read-only, in-memory cache that serves consistent read queries and does not accept DML; write operations continue to go to the primary database. Option A is incorrect because True Cache does not replace Data Guard — Data Guard remains Oracle's disaster-recovery and data-protection solution, and True Cache is complementary to it. Option B is incorrect because True Cache does not require a dedicated database; it is an additional cache instance associated with the primary database, not a separate standalone database deployment. Option C is incorrect because True Cache does not support DML operations; it only serves read-only queries, with all DML directed to the primary database.

Answer

It replaces Data Guard

Answer

It requires a dedicated database

Answer

It supports DML operations

Question 3

Which THREE of the following are components of AI Vector Search in 23ai?

Accepted Answer

HNSW indexing. Option A (HNSW indexing) is correct because Oracle AI Vector Search in 23ai supports Hierarchical Navigable Small World (HNSW) vector indexes, an in-memory graph-based index type used for fast approximate nearest-neighbor similarity searches on vector embeddings. Option C (Vector distance functions) is correct because AI Vector Search provides built-in distance metrics such as COSINE, DOT (inner product), EUCLIDEAN, and EUCLIDEAN_SQUARED, exposed through SQL functions like VECTOR_DISTANCE, which are essential for ranking vectors by similarity. Option D (VECTOR data type) is correct because 23ai introduces the native VECTOR data type, allowing vector embeddings to be stored directly in tables as a first-class column type with a specified number of dimensions and format (e.g., FLOAT32, INT8, BINARY). Option B (Automatic backup) is not a component of AI Vector Search; it is a general database administration feature provided by RMAN and Oracle Cloud infrastructure, unrelated to vector search functionality. Option E (JSON Schema validation) is not part of AI Vector Search either; JSON schema validation relates to Oracle's JSON capabilities (e.g., IS JSON VALIDATE constraints), not to vector indexing, storage, or similarity search.

Answer

Automatic backup

Answer

JSON Schema validation

Question 4

In 23ai, when using the 'JSON relational duality' feature, how are changes to the JSON document propagated to the underlying tables?

Accepted Answer

Automatically propagated by the database. JSON Relational Duality in Oracle 23ai provides automatic bidirectional synchronization between JSON documents and their underlying relational tables. When a JSON document is updated through a duality view, the database automatically propagates the changes to the base tables using the defined mappings. No manual re-synchronization, triggers, or stored procedures are required.

Answer

Only allowed via stored procedures

Answer

Manual re-synchronization

Answer

Requires a trigger

Question 5

You want to use Memoptimized Write to speed up data ingestion. Which type of table should you use?

Accepted Answer

Heap-organized table. Memoptimized Write is a feature introduced in Oracle Database 18c that accelerates high-volume INSERT operations by buffering inserts in the SGA and writing them directly to the database using a fast path, bypassing much of the standard buffer cache and redo overhead. It is only supported on heap-organized tables, since these tables have a simple rowid-based physical structure that the fast-path insert mechanism can target. Index-organized tables, external tables, and global temporary tables do not support the Memoptimized Write clause.

Answer

External table

Answer

Global temporary table

Answer

Index-organized table

Question 6

What happens if a user tries to drop a Blockchain table that still contains rows within the retention period?

Accepted Answer

The command fails with an error. Blockchain tables in Oracle enforce immutability for rows within the retention period. If a user attempts to DROP a blockchain table that still holds rows whose retention window has not expired, Oracle raises an error (ORA-05723 or similar) and the DROP is rejected. This protects the tamper-evident ledger guarantee — rows cannot be removed before their retention period elapses.

Answer

The command succeeds

Answer

The table is renamed to _OLD

Answer

The rows are archived then dropped

Question 7

When using Wide Tables, what is the impact on performance for queries that select only a small subset of columns?

Accepted Answer

No significant performance degradation. Wide Tables in 23ai are optimized to handle many columns without significant performance degradation compared to vertical splitting.

Answer

Severe performance degradation

Answer

Requires partitioning

Answer

Requires a full table scan

Question 8

For Memoptimized Write, where is the data initially stored before being written to the database?

Accepted Answer

SGA memory buffer. Memoptimized Write is an Oracle Database feature that accelerates high-frequency, low-latency inserts by buffering incoming row data in a dedicated area of the SGA called the Memoptimized Rowstore. Once the buffer fills or a threshold is reached, the data is asynchronously written to the database in bulk, reducing commit overhead and improving throughput. The SGA buffer acts as a staging area, so the data is not immediately persisted to disk or redo logs until the write is flushed.

Answer

Flash cache

Answer

Undo tablespace

Answer

Redo log

Question 9

Which TWO features of 23ai help improve performance of high-concurrency workloads?

Accepted Answer

Lock-Free Reservations. Lock-Free Reservations (B) is correct because it lets concurrent transactions reserve a column value (e.g., inventory quantity) without holding a row lock, eliminating lock contention and enabling high-concurrency OLTP throughput. Memoptimized Write (E) is correct because it uses the Memoptimized Rowstore fast ingest path to buffer inserts in memory and write them to disk asynchronously, dramatically improving insert performance under many concurrent sessions. Wide tables (A) merely allow up to 1,000 columns and do not by themselves improve concurrency. Blockchain tables (C) add tamper-evident row chaining for auditing, which adds overhead rather than boosting concurrency. JSON Duality (D) provides dual relational/JSON access to the same data for developer flexibility, not a high-concurrency performance feature.

Answer

Wide tables

Answer

Blockchain tables

Answer

JSON Duality

Question 10

What is the main requirement for an existing table to be converted into a Blockchain table?

Accepted Answer

Tables cannot be converted to Blockchain. Existing tables cannot be converted into Blockchain tables. Blockchain tables must be created as such from the start using the CREATE BLOCKCHAIN TABLE statement. There is no ALTER TABLE command to convert a regular table into a blockchain table.

Answer

The table must have an index

Answer

The table must be partitioned

Answer

The table must be empty

Question 11

In True Cache, how is consistency maintained with the primary database?

Accepted Answer

Asynchronous replication. True Cache maintains consistency with the primary database through asynchronous replication of redo/change data. The True Cache instance applies changes from the primary's redo stream in near real-time, so it is not transactionally synchronous but is kept current with low lag. This design lets True Cache serve read-only queries without blocking on the primary.

Answer

Manual refreshes

Answer

Polling

Answer

Synchronous writes

Question 12

When using the VECTOR data type, what is the maximum number of dimensions supported?

Accepted Answer

2048. Oracle 23ai supports up to 2048 dimensions for the vector data type.

Answer

512

Answer

1024

Answer

4096

Question 13

Which type of operation is ideal for Blockchain tables?

Accepted Answer

Audit logs and financial records. Blockchain tables are designed for insert-only, tamper-evident records where each row is cryptographically hashed and chained to the previous row. Audit logs and financial records are the ideal use case because they require an immutable, verifiable history that can be used for compliance and auditing. These tables prevent updates and deletions, ensuring data integrity and non-repudiation.

Answer

General configuration tables

Answer

Transient cache data

Answer

High-frequency temporary storage

Question 14

Which THREE operations are permitted on Blockchain tables?

Accepted Answer

VERIFY. Blockchain tables in Oracle Database are append-only, tamper-evident tables, so the permitted operations are INSERT (C), SELECT (E), and VERIFY (A). INSERT (C) is allowed because new rows can only be added to the end of the chain, which is how the table grows while preserving its hash-linked row history. SELECT (E) is allowed because reading and querying rows is unrestricted and does not alter the chain. VERIFY (A) is allowed because it is the special operation that checks the table's hashes and signatures to confirm no rows have been tampered with. DELETE (B) and UPDATE (D) are not permitted, since blockchain tables are append-only and rows cannot be modified or removed after insertion.

Answer

DELETE

Answer

UPDATE

Question 15

What is the primary function of the 'JSON relational duality' feature?

Accepted Answer

To map relational data to JSON. It provides a unified way to work with JSON and relational data simultaneously.

Answer

To compress JSON data

Answer

To encrypt JSON data

Answer

To index JSON data