Cosmos DB — Animate | AI Wisdom

← Back to topic
🎞 ANIMATE · Mode 3 of 5 · +100 XP— Azure Cosmos DB
📋Container
⚙️Partition
🔄Point Read
✅Query
🎯 Next: Test →
Step 1 / 4
Speed
Create Container
Container with partitionKey "/customerId" — data distributed by customer💡 Analogy
Like a post office with mailboxes sorted by customer — all your mail is in one spot.
Cosmos DB stores data in containers with a partition key. All documents with the same partition key are stored together — enabling fast per-customer queries.
Console Output
> Container "orders" created
> Partition key: /customerId
> TTL: 30 days (automatic cleanup)
> Throughput: 400 RU/s (serverless)
●cosmos_db.ts
Create ContainerL8–13
1import { CosmosClient, PartitionKey } from '@azure/cosmos';
2 
3const client = new CosmosClient({
4  endpoint: process.env.COSMOS_ENDPOINT!,
5  key: process.env.COSMOS_KEY!,
6});
7 
8const { database } = await client.databases.createIfNotExists({ id: 'ecommerce' });
9const { container } = await database.containers.createIfNotExists({
10  id: 'orders',
11  partitionKey: { kind: 'Hash', paths: ['/customerId'] },
12  defaultTtl: 86400 * 30,  // 30-day TTL on documents
13});
14 
15// Create — partition key is required
16await container.items.create({
17  id: 'order-001',
18  customerId: 'cust-123',      // partition key value
19  items: [{ productId: 'p1', qty: 2, price: 29.99 }],
20  status: 'pending',
21  createdAt: new Date().toISOString(),
22});
23 
24// Read — provide partition key for efficient point read (RU cost: 1)
25const { resource: order } = await container
26  .item('order-001', 'cust-123')    // id + partition key
27  .read();
28 
29// Query — cross-partition if partition key not in WHERE
30const { resources: orders } = await container.items.query({
31  query: 'SELECT * FROM c WHERE c.customerId = @id AND c.status = @status',
32  parameters: [
33    { name: '@id', value: 'cust-123' },
34    { name: '@status', value: 'pending' },
35  ],
36}).fetchAll();
36 lines
Timeline — 4 Steps
Create Container
Container with partitionKey "/customerId" — data distributed by customer💡 Analogy
Like a post office with mailboxes sorted by customer — all your mail is in one spot.
Cosmos DB stores data in containers with a partition key. All documents with the same partition key are stored together — enabling fast per-customer queries.
Console Output
> Container "orders" created
> Partition key: /customerId
> TTL: 30 days (automatic cleanup)
> Throughput: 400 RU/s (serverless)

Was this page helpful?

Discussion

Container with partitionKey "/customerId" — data distributed by customer

💡 Analogy

Like a post office with mailboxes sorted by customer — all your mail is in one spot.

Cosmos DB stores data in containers with a partition key. All documents with the same partition key are stored together — enabling fast per-customer queries.

Console Output

> Container "orders" created
> Partition key: /customerId
> TTL: 30 days (automatic cleanup)
> Throughput: 400 RU/s (serverless)

Container with partitionKey "/customerId" — data distributed by customer

💡 Analogy

Like a post office with mailboxes sorted by customer — all your mail is in one spot.

Cosmos DB stores data in containers with a partition key. All documents with the same partition key are stored together — enabling fast per-customer queries.

Console Output

> Container "orders" created
> Partition key: /customerId
> TTL: 30 days (automatic cleanup)
> Throughput: 400 RU/s (serverless)