How to Model Versioned Data in DynamoDB

Versioning is a common requirement in any modern app.

Some examples can include tracking changes to user profiles or storing document revisions. 

Versioning data enables traceability, rollback, and audit compliance capabilities, which are often very useful.

In this article, let’s explore how we can model versioned data efficiently with real-world strategies and tips.

Why Versioning is Challenging in DynamoDB

DynamoDB’s schema-less design and lack of relational features mean there’s no out-of-the-box versioning like in SQL (e.g., triggers or temporal tables). 

But with the right primary key structure and access patterns, you can build powerful versioning systems that scale.

Core Design Pattern: Primary Key Versioning

The most common way to model versioned data in DynamoDB is by storing each version as a new item in the same table.

Here’s the general pattern:

• Partition key: the ID of the entity (e.g. doc#123)

• Sort Key: a prefix of the version number (e.g. version#001)

Here’s an example data model that tracks document versions:

This structure supports:

• Getting the full version history: query pk = “doc#123”

• Getting the latest version: query with ScanIndexForward: false, Limit: 1

• Reverting to prior version: copy the old version back as the new one

Here, versioning just becomes another access pattern.

Alternative Data Model: Timestamp-based Sort Key

For use cases where versions are based on timestamps, use an ISO timestamp as the sort key:

• Partition key: doc#123

• Sort key: v#2025–01–01T12:00:00:00Z#user#101

This approach is especially useful in audit logs or time-series history where natural ordering is key.

Just make sure timestamps are globally unique to prevent overwrites (using the userID helps with that).

Handling Rollbacks

Reverting to a previous version is simple. Read and copy the version you want to revert to, increment the version number, and write a new item with that new version number.

This preserves the auditability without deleting or mutating the history.

Deletion and TTL

To avoid infinite accumulation of versions, especially in the context of costs, you can add TTL attributes to older versions.

You can keep the latest versions and apply a TTL to delete older items after X days.

Use a query to list all versions to find versions to TTL periodically.

Summary

Modeling versioned data in DynamoDB requires careful key design due to its schema-less nature. 

We discussed strategies like primary key versioning and timestamp-based sort keys offer scalable solutions. 

With these patterns, you can efficiently manage data history, enable rollbacks, and maintain audit compliance while controlling storage costs with TTL policies.

👋 My name is Uriel Bitton and I hope you learned something in this edition of Excelling With DynamoDB.

📅 If you're looking for help with DynamoDB, let's have a quick chat.

🙌 I hope to see you in next week's edition!