Question

A pricing service needs to fetch 500 specific product documents by their known id and known partition key value (category) in a single operation. The developer initially wrote a query using WHERE c.id IN (...) that fans out across many logical partitions and consumes high RU. Using the .NET SDK, which approach retrieves these known items most RU-efficiently?

Accepted Answer

Use the SDK's ReadManyItemsAsync (read-many) API, passing a list of (id, partition key) tuples so each item is retrieved as an optimized point-read batch.

Answer

Wrap the IN query inside a stored procedure so it executes server-side within a single transaction.

Answer

Enable autoscale throughput so the higher RU consumption of the cross-partition query is automatically absorbed.

Answer

Switch the connection mode from direct to gateway to reduce the RU charge of the fan-out query.