Application Programming Interfaces APIs

Now that we’ve covered why Azure Cosmos DB is an outstanding semi-structured data platform and how to provision a Cosmos DB account, this article dives into the five supported APIs. You’ll learn how and when to use the native NoSQL (Core) API, MongoDB, Cassandra, Gremlin (graph), and Table Storage. By the end, you’ll be able to pick the right interface for your workload.

Two core advantages differentiate Cosmos DB from Azure Storage Accounts:

Global distribution & geo-redundancy
Deploy multiple writable regions worldwide for low-latency reads and writes with built-in replication.
Infinite throughput scalability
Purchase as many Request Units (RUs) per second as you need—up to tens of millions—without the fixed limits of Azure Storage.

NoSQL (Core) API

The native NoSQL API stores JSON documents and offers a familiar SQL-like query language for JSON data. If you’re coming from relational databases, you can leverage your SQL skills to query documents.

The image is an illustration of Cosmos DB's NoSQL API, highlighting its use of JSON documents and a SQL-like language for querying.

Key features:

JSON-document model with automatic indexing
SQL-style queries: SELECT * FROM c WHERE c.category = 'electronics'
Cross-partition queries with rich filtering and aggregation

Note

The Core API automatically indexes all JSON properties by default, so there’s no need to manage schema migrations.

MongoDB API

Azure Cosmos DB’s MongoDB API is wire-protocol compatible with existing MongoDB drivers. Simply point your application at Cosmos DB—no code changes required. You can also enforce schemas using JSON Schema validation.

The image is a diagram showing MongoDB's integration with Cosmos DB, highlighting its document-oriented, NoSQL database features using JSON-like documents. It mentions optional schema use for structured tables and compatibility with non-Azure implementations.

Key points:

Full support for MongoDB wire protocol v3.2+
Zero-downtime migration with the Azure Cosmos DB Data Migration tool
Optional JSON Schema enforcement for stricter models

Note

You can leverage MongoDB tools like mongoexport or mongodump to migrate data directly into Cosmos DB.

Cassandra API

Built for wide-column workloads, the Cassandra API lets you define column families to group related columns for fast reads and writes. It’s wire-compatible with existing Cassandra drivers.

The image is an infographic about Cassandra, a column-oriented NoSQL database, highlighting its focus on columns, wide column speed, column family organization, and use of Cassandra Query Language (CQL). It also mentions integration with Cosmos DB.

Highlights:

Uses Cassandra Query Language (CQL) for familiar syntax
Support for existing Cassandra driver versions
Integrate with Apache Spark or Azure Databricks for ETL and analytics

Gremlin (Graph) API

The Gremlin API enables graph databases where both vertices (nodes) and edges (relationships) can store properties. It excels when you need to traverse complex relationships.

The image is an illustration of Apache Gremlin API for Cosmos DB, highlighting its use for working with graph or network databases, where both vertices and edges can hold data. It emphasizes the importance of relationships between items.

Common scenarios:

Social networks (friends, followers)
Network topology mapping (devices, connections)
Fraud detection with path-finding queries

Table Storage API

If you need a simple key-value store with schema-less tables, the Table Storage API gives you the same programming model as Azure Storage Tables—plus Cosmos DB’s global distribution and unlimited RUs.

Key advantages:

PartitionKey/RowKey indexing for fast lookups
Automatic horizontal partitioning and replication
Ideal for telemetry, session stores, and lightweight metadata

API Comparison

API	Data Model	Query Language	Best Use Case
Core (NoSQL)	JSON documents	SQL-like	Flexible JSON workloads with rich querying
MongoDB	JSON documents	MongoDB Query API	Existing MongoDB apps with minimal changes
Cassandra	Wide-column	CQL	High-throughput columnar reads/writes
Gremlin (Graph)	Graph	Gremlin traversal	Complex relationship queries and traversals
Table Storage	Key-value	OData / Table query	Simple schema-less tables with fast lookups

Links and References

Watch Video

Watch video content