Bedrock Knowledge Bases

Currently, Amazon Bedrock Knowledge Bases does not have a built-in feature specifically for storing page numbers from processed CSV files in the metadata of embeddings. However, the service does offer metadata customization capabilities that could potentially be used to achieve a similar result.

When ingesting CSV files, Bedrock Knowledge Bases allows you to designate specific columns as content fields and others as metadata fields. This feature enables more efficient data management, especially for large CSV file datasets. You can provide metadata files alongside the source CSV files to specify which fields should be treated as content and which as metadata.

If your CSV contains a column with page number information, you could potentially include this as part of the metadata during the ingestion process. To do this, you would need to create a metadata file (with a .csv.metadata.json suffix) that specifies the page number field as one of the metadata fields to include.

While this approach doesn't directly store "page numbers" in the traditional sense (as CSVs don't have pages like PDFs do), it does allow you to associate a numerical identifier or sequence number with each row of data, which could serve a similar purpose in terms of referencing specific parts of your dataset.

It's worth noting that Bedrock Knowledge Bases is continually evolving, and AWS regularly adds new features. If precise page number storage for CSVs is crucial for your use case, it may be worth reaching out to AWS support or checking their documentation for any updates on this specific functionality.
Sources
Amazon Bedrock Knowledge Bases now supports advanced parsing, chunking, and query reformulation giving greater control of accuracy in RAG based applications | AWS Machine Learning Blog
Amazon Bedrock Knowledge Bases now supports advanced parsing, chunking, and query reformulation giving greater control of accuracy in RAG based applications | AWS Machine Learning Blog