Answer-first summary for fast verification
Answer: Include metadata columns for article date and topic to support metadata filtering.
Option B is correct because it uses metadata filtering, which is the most efficient and minimal effort approach for this use case. Databricks Vector Search supports metadata columns that can be used for filtering during search operations. By including article date and topic as metadata columns, the engineer can perform semantic similarity search while filtering results to articles within 10 days of the specified date and matching the relevant topic. This requires no additional infrastructure or complex logic. Option A creates unnecessary complexity with separate indexes and classifier models. Option C ignores the date filtering requirement entirely. Option D uses an inefficient manual partitioning approach that doesn't leverage Databricks Vector Search's built-in capabilities.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
How can a Generative AI Engineer configure a Databricks Vector Search index with minimal effort to support querying news articles by topic while also filtering results to within 10 days of a specified date, such as for the query "Tell me about monster truck news around January 5th 1992"?
A
Create separate indexes by topic and add a classifier model to appropriately pick the best index.
B
Include metadata columns for article date and topic to support metadata filtering.
C
Pass the query directly to the vector search index and return the best articles.
D
Split articles by 10 day blocks and return the block closest to the query.
No comments yet.