Amazon DynamoDB is widely known for its speed, scalability, and ability to handle massive workloads without compromising performance. One of its most powerful features is the Global Secondary Index (GSI), which allows developers to query data efficiently using alternate keys.

Traditionally, GSIs were limited to a single partition key and a single sort key. However, with the introduction of multi-key support, DynamoDB takes flexibility and performance to the next level. Now DynamoDB supports up to four attributes for the partition key (PK) and four attributes for the sort key (SK) in a composite manner to identify items as part of a GSI.

Multi-key support enables you to define multiple attributes as keys within a single GSI, allowing queries to span across different dimensions without creating multiple indexes. This means you can:

• Reduce the number of GSIs needed for complex queries.
• Optimise storage and cost by consolidating indexes.
• Improve query performance for diverse access patterns.

Why Does This Matter?

In real-world applications, data access patterns evolve. For example:

• IoT systems might query device reading data by location, year, month or timestamp.
• E-commerce platforms may need to query orders by customer ID, product category, or status.

Previously, you had to create separate GSIs for each pattern, which increased complexity and cost. Multi-key support simplifies this by allowing one GSI to serve multiple query dimensions.

Key Benefits

• Cost Efficiency: Fewer GSIs mean lower storage and maintenance costs.
• Simplified Design: Reduce schema complexity and operational overhead.
• Improved Query Flexibility: Support multiple access patterns without redesigning your database.

Getting Started

To leverage multi-key support for time-series data with hierarchical time attributes (year → month → day → timestamp) for efficient queries at any time granularity without date parsing or manipulation.

Base Table Design:

Figure 1:  DynamoDB table name – IoTReadings

Indexed Table – GSI with multiple key attributes.

Figure 2: Secondary Index name – DeviceLocationTimeIndex

To create a base table and its GSI, create a JSON file named “table.json” shown below:

To create a table using AWS CLI, run the following command shown below.

Added a set of items to a table. Below are sample items added into the table as shown in Figure 3, using the AWS CLI command shown below.

Figure 3: Items added to the table

Access Pattern Requirement:

1. All readings for a device in a specific location
2. Readings for the device in a specific location for a specific year and specific month

Similarly, you may have various requirements that you can fulfil with GSI for reading data related to a specific device, with granularity down to a specific day and time.

To retrieve the data for the access pattern requirement 1, run the command below using the AWS CLI, and the output is shown in Figure 4 below.

Output of access pattern requirement 1, displaying readings of device – D001, in location – L001.

Figure 4: output of query operation on GSI for access requirement 1

To retrieve the data for the access pattern requirement 2, run the command below using AWS CLI.

Output of access pattern requirement 1, displaying readings of device – D001, in location – L001, for year 2024 and month 01.

Figure 5: output of query operation on GSI for access requirement 2

For detailed implementation guidance, check the official instructions on using global secondary indexes in DynamoDB, or to know more about DynamoDB or other NoSQL solutions on AWS, you can also check out courses like “Build Modern Applications with AWS NoSQL Databases”.

Flexible DynamoDB Querying

Multi-key support for GSIs in DynamoDB is a significant enhancement for developers building scalable, flexible applications. It reduces complexity, optimises cost, and unlocks new possibilities for data access patterns, all while maintaining DynamoDB’s hallmark performance. This enhancement is a major leap forward for anyone architecting scalable, query-efficient NoSQL solutions on AWS.

About CloudThat