cosmos-repository

Cosmos DB pagination can be implemented using a FeedIterator to efficiently stream large result sets without loading everything into memory. This approach standardizes repository patterns alongside LINQ queries and discriminator filtering.

Using LINQ for Cosmos DB data access involves GetItemLinqQueryable to apply discriminator filters directly against containers. This standardizes query patterns and reduces boilerplate code in .NET microservices.

FeedIterator streams paginated results to control memory usage, while point reads target specific items by ID and partition key. Tracking Request Units during these operations helps optimize Cosmos DB performance and cost.

ETags enable safe upserts in Cosmos DB by managing optimistic concurrency control. Checking the ETag during repository updates reduces conflicts in concurrent environments and prevents accidental overwrites.

This repository pattern is designed for .NET microservices implementing Cosmos DB data access. It provides a reusable base for point reads, LINQ queries, and FeedIterator pagination within distributed service architectures.

Loading too much data into memory happens when querying without FeedIterator pagination. Implementing FeedIterator streams large result sets efficiently, applying discriminator filters to limit the queried container data.