[FEATURE REQ] Enhance findByIds to Use readMany for Partition Key Optimized Queries or Provide Direct Support for readMany in Spring Data Cosmos SDK

[aanaji]

In Cosmos DB, the readMany API provides a more efficient way to fetch multiple items when their id values align with the partition key. However, in the Spring Data Cosmos SDK, the findByIds method does not currently leverage this optimisation. Supporting readMany in findByIds or providing explicit support for readMany in the SDK would significantly improve performance and reduce RU consumption for such use cases.

Proposed Solution:

1. Enhance findByIds:

   • Automatically detect when the id parameter matches the partition key.
   • Use the readMany API internally in such cases to optimise performance.

2. Direct Support for readMany:

   • Provide a new method, e.g., readManyByIds, in the Spring Data Cosmos SDK to allow developers to explicitly utilize the readMany API.
   • This method could accept a list of IDs and their corresponding partition keys to fetch the documents efficiently.

Use Case:

• Scenario: Fetching multiple records by their IDs where id is also the partition key.
• Benefits:
  • Reduces the number of round-trips to Cosmos DB.
  • Lowers RU consumption compared to fetching items individually.

Expected Behavior:

• When id matches the partition key, findByIds should leverage the readMany API.
• For the readManyByIds method, developers can explicitly specify the usage of readMany.

References:

• https://devblogs.microsoft.com/cosmosdb/read-many-items-fast-with-the-java-sdk-for-azure-cosmos-db/

Impact:

• Improves performance and reduces RU costs in high-throughput applications.

[github-actions]

Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @kushagraThapar @pjohari-ms @TheovanKraay.

[kushagraThapar]

@trande4884 can you please take a look at this, thanks!