NoSQL databases are versatile, offering a variety of data models. Let’s go through four prominent types of NoSQL databases and look at examples of each.

Key-Value Stores

In this type, data is stored as key-value pairs. It’s a simple and fast option, suitable for tasks like caching.

Example:

• Database: Amazon DynamoDB, Redis.
• In TypeScript:

  // DynamoDB:  dynamoDB.put({TableName: 'userTable', Item: { id: { N: '123' }, name: { S: 'Alice' }}});
  // Redis: redisClient.set('userID:123', 'Alice');
  // or retrieve: redisClient.get('userID:123', (err, reply) => { console.log(reply); });

Wide-Column Stores

Wide-column stores use column families to group related data. Individual records don’t need to have the same columns. This structure is ideal for analytical workloads.

Example:

• Database: Google Bigtable, Apache Cassandra.
• In TypeScript:

  // Bigtable: bigtableInstance.table('your-table').row('row-key').save({ columnFamily: { columnQualifier: 'columnValues' } });
  // Cassandra: session.execute("INSERT INTO users (id, name, age) VALUES (123, 'Alice', 30)");

Document Stores

These databases store each record as a document (often in JSON or BSON format), with a unique identifier. They are preferred for content management systems and real-time analytics.

Example:

• Database: MongoDB, Couchbase.
• In TypeScript:

  // MongoDB: db.collection('users').insertOne({ _id: 123, name: 'Alice', age: 30 });
  // Couchbase: bucket.upsert('user::123', { name: 'Alice', age: 30 });

Graph Databases

These are ideal for data with complex relationships. Instead of tables or collections, they use nodes, edges, and properties to represent relational data.

Example:

• Database: Neo4j, Amazon Neptune.
• In TypeScript:

  // Neo4j: cypherQuery('CREATE (a:Person {name: "Alice"})-[:LIKES]->(b:Person {name: "Bob"})');
  // Neptune: neptune.think("I think Alice likes Bob");