Question: What is an access pattern in DynamoDB?

Answer

An access pattern is one specific way your application reads or writes data, phrased as a concrete request: "get a customer by ID", "list a customer's orders from the last 30 days, newest first" or "find open orders for a warehouse". In DynamoDB you list your access patterns before you design the table, because each one should map to an efficient key-based operation:

• GetItem with the full primary key, or
• Query on a partition key, optionally narrowed by a sort key condition, against the table or a secondary index.

If neither a primary key nor an index supports the lookup, consider changing the model or querying an exported copy. A Scan evaluates items across the table and is often expensive for repeated application lookups. That's why the relational habit of modeling entities first and writing queries later is reversed: in DynamoDB, the queries shape the keys.

Example for a small shop, using one table with generic PK and SK attributes:

1. Get a customer profile: GetItem with PK = CUSTOMER#123 and SK = PROFILE.
2. List a customer's orders, newest first: Query with PK = CUSTOMER#123 and SK beginning with ORDER#, where order sort keys look like ORDER#2026-10-23#987, and ScanIndexForward: false.
3. Get an order by ID: Query a GSI keyed on orderId.
4. List a warehouse's open orders by date: Query a sparse GSI keyed on openWarehouse and createdAt, attributes you set only while an order is open.

For each pattern, also note how often it runs and how fast it must be. A rare reporting query may not deserve an index at all; it can run against an export instead.

Tools help here. NoSQL Workbench supports data modeling, and Dynobase's Single-Table Designer lets you model entities, access patterns and GSIs visually. Worked examples are in the single-table design examples library.

Source: AWS DynamoDB documentation.