1Z0-184-25 Question 35
Single answerYou are working with an OCI-hosted Oracle Database table that stores product embeddings in a vector-type column of dimension 128. Your data science team has decided to expand these embeddings to dimension 256 to improve recommendation accuracy. You want to accomplish this without losing any existing data. What is the recommended approach?
- A
Use �ALTER TABLE ... MODIFY� on the existing vector column to change its dimension from 128 to 256 in place.
- B
Delete all existing data, then use �ALTER TABLE ... ADD� to create a new 256-dimensional vector column and reload the data.
- C
Add a new 256-dimensional vector column using �ALTER TABLE ... ADD� and migrate the existing embeddings before dropping the old column.
- D
Disable the current vector column, create an invisible column with the new dimension, and then re-enable the original column with the updated dimension.
Show answer and explanation
Correct answer: C
Explanation
DDL operations on vector columns in Oracle Database require careful planning, especially when the vector dimension must be altered. Because you cannot simply modify a vector column� dimension, the best practice is to create a new vector column with the required dimension and migrate data to it before dropping the old column. This approach aligns with Oracle's guidelines for altering complex data types. Refer to Oracle Database documentation on vector data types and schema evolution for more details.
- A. Incorrect.
Option 1 is incorrect. Oracle Database does not allow an in-place dimensional expansion of a vector column with �ALTER TABLE ... MODIFY� for the vector data type. You cannot simply change its dimension without other steps to handle existing data.
- B. Incorrect.
Option 2 is incorrect. While it would work, it would require deleting and reloading your data, which risks data loss or downtime. It is not the recommended approach when you can migrate data in-place instead.
- C. Correct.
Option 3 is correct. The recommended approach is to add a new vector column with the desired dimension, copy or transform the existing data into that new column, and then drop the old column once the migration is successful. This preserves data while allowing for a smooth schema evolution.
- D. Incorrect.
Option 4 is incorrect. There is no straightforward mechanism to �disable� a vector column or directly re-enable it with a different dimension. Making the column invisible only hides it from queries but does not change its data type or dimension.