1Z0-184-25 Question 11
Select 2Your data science team is building a recommendation engine in Oracle Cloud Infrastructure� Autonomous Database. They want to store product embeddings as vectors and perform semantic similarity searches efficiently. Which two steps should you take to enable vector-based semantic queries on these embeddings?
- A
Create a new column of a vector data type and store the product embeddings in that column.
- B
Populate the vector column with embedding data and build an approximate nearest neighbor (ANN) index on the column.
- C
Embed all product data into a single BLOB column and rely on standard SQL text indexing to perform similarity searches.
- D
Use an external key-value store outside Autonomous Database and link it back to the main tables through custom code.
Show answer and explanation
Correct answers: A, B
Explanation
To enable efficient semantic queries on embeddings in Oracle Autonomous Database, you must use the native vector data type and create a vector index (typically an approximate nearest neighbor index) to accelerate the similarity searches. These features are designed to handle high-dimensional vector data, making them ideal for recommendation engines, semantic search, and other machine learning workloads. Refer to Oracle� official documentation on Autonomous Database vector support for details on creating vector columns and ANN indexes.
- A. Correct.
Option 1 is correct. Oracle Database provides a native vector data type for storing embeddings. By creating a column of this vector type, you can store and manage embeddings in a structured, optimized manner.
- B. Correct.
Option 2 is correct. In addition to storing embeddings in a vector column, creating an approximate nearest neighbor (ANN) index is essential for efficient semantic queries. This index is specifically designed to accelerate similarity searches on vector data.
- C. Incorrect.
Option 3 is incorrect. Storing embeddings in a BLOB column and relying on standard text indexes will not provide the specialized search capabilities needed for semantic queries on vector data. Text indexes are optimized for keyword matching, not vector similarity.
- D. Incorrect.
Option 4 is incorrect. Leveraging an external key-value store complicates the architecture and does not integrate natively with Autonomous Database� vector data type and indexing features. It increases maintenance overhead and provides no direct performance benefit for vector similarity searches.