DAS-C01 · Question #14
A company has 1 million scanned documents stored as image files in Amazon S3. The documents contain typewritten application forms with information including the applicant first name, applicant last…
The correct answer is B. Index the metadata and the Amazon S3 location of the image file in Amazon Elasticsearch. To enable data analysts to quickly search and analyze extracted metadata from 1 million scanned documents in S3, prioritizing query performance, indexing the metadata and the S3 image location in Amazon Elasticsearch Service is the optimal solution. Elasticsearch provides fast…
Question
A company has 1 million scanned documents stored as image files in Amazon S3. The documents contain typewritten application forms with information including the applicant first name, applicant last name, application date, application type, and application text. The company has developed a machine learning algorithm to extract the metadata values from the scanned documents. The company wants to allow internal data analysts to analyze and find applications using the applicant name, application date, or application text. The original images should also be downloadable. Cost control is secondary to query performance. Which solution organizes the images and metadata to drive insights while meeting the requirements?
Options
- AFor each image, use object tags to add the metadata.
- BIndex the metadata and the Amazon S3 location of the image file in Amazon Elasticsearch
- CStore the metadata and the Amazon S3 location of the image file in an Amazon Redshift table.
- DStore the metadata and the Amazon S3 location of the image files in an Apache Parquet file in
How the community answered
(34 responses)- A15% (5)
- B76% (26)
- C6% (2)
- D3% (1)
Why each option
To enable data analysts to quickly search and analyze extracted metadata from 1 million scanned documents in S3, prioritizing query performance, indexing the metadata and the S3 image location in Amazon Elasticsearch Service is the optimal solution. Elasticsearch provides fast full-text search and analytical capabilities, and stores the S3 path for image downloads.
Amazon S3 object tags are limited to 10 key-value pairs per object and are not designed for complex searching, full-text queries, or large volumes of metadata. This approach would quickly become unmanageable and perform poorly for the specified analytical requirements.
Amazon Elasticsearch Service (now OpenSearch Service) is specifically designed for full-text search, complex queries, and analytics, making it highly performant for finding applications using applicant names, dates, or application text. By indexing the extracted metadata along with the Amazon S3 location of the image files, analysts can perform fast searches and then easily access the original documents, fulfilling the query performance requirement.
Amazon Redshift is a data warehouse optimized for complex analytical queries and aggregations over large datasets, not for low-latency, full-text search queries like 'find applications' using specific text fields. While it can store the data, its performance for this search-oriented workload would be inferior to Elasticsearch.
Storing metadata in Apache Parquet files in Amazon S3, even with Amazon Athena for querying, is optimized for large-scale analytical scans and aggregations. It is not designed for the low-latency, full-text search, and ad-hoc item retrieval required to efficiently 'find applications' by specific text fields, where Elasticsearch excels.
Concept tested: Search and analytics on metadata using AWS OpenSearch
Source: https://aws.amazon.com/opensearch-service/
Topics
Community Discussion
No community discussion yet for this question.