Amazon Quick says goodbye to manual pre-joins with multi-dataset topics!

Hi, it's Shii! Today I found an update to AWS's BI service, Amazon Quick. It might sound a bit understated, but it's a big deal if you work with multiple datasets, so let me walk you through it!

AWS What's New aws.amazon.com

What was announced?

Over on AWS What's New, AWS announced that Amazon Quick now supports multi-dataset topics. You can model the relationships across multiple datasets inside a single topic, and then use that model to build dashboards or answer questions in natural language.

The story so far

Up until now, answering a question that spanned multiple datasets in a visual meant pre-joining the data into a single dataset first. That came with a few pain points:

• You had to write manual JOIN logic during data preparation
• The joined dataset consumed extra SPICE capacity
• You had to rebuild datasets whenever the use case or the model changed

In other words, every new analytical angle meant another dataset to manage behind the scenes.

What changes

With multi-dataset topics, a topic itself becomes a reusable relational data model. You add multiple datasets, define the relationships once, and Quick performs the joins for you at query time.

This works for both dashboard building and natural-language Q&A.

• Dashboard building: The topic acts as the data model, so a single visual can pull fields from multiple datasets. Quick generates the underlying join automatically, so you author the semantic model once and reuse it across every visual in the dashboard
• Natural-language Q&A: You can point a chat agent directly at a topic and ask questions. The agent reads the relationships defined in the topic and performs the joins across datasets at runtime, so there's no need to pre-join tables or prepare data first

Because both the dashboard and the chat agent query the same topic, one governed semantic model becomes the single source of truth for both people and agents.

Dive Deep

Multi-dataset topics carry over your existing dataset permissions as-is. They also support row-level security (RLS) and column-level security (CLS), so the governance you already set up flows through to every cross-dataset visual and every chat agent answer. You don't have to redo your access controls.

Multi-dataset topics are now generally available in every AWS Region where Amazon Quick is available. For a walkthrough on getting started, AWS published a detailed post on the AWS Machine Learning Blog.

Wrap-up

• Amazon Quick now supports multi-dataset topics, letting you model relationships across multiple datasets in a single topic
• Manual pre-joins, extra SPICE capacity, and rebuilding datasets are no longer required
• Both dashboards and natural-language Q&A can query the same topic and get automatic joins at runtime
• Existing dataset permissions and RLS/CLS carry over automatically
• Now generally available in all AWS Regions

This one is especially worth checking out if you've been juggling multiple datasets and dealing with the overhead of managing joins between them!