DynamoDB Books Worth Reading in 2026

There are two DynamoDB books worth paying for, and they are not substitutes. One teaches you how to think about modeling; the other documents the service. Most "best DynamoDB books" lists rank them without mentioning the thing that decides which to buy: both were published before features you are probably using today. This page says which parts have aged and which have not.

Which one you want

The DynamoDB Book by Alex DeBrie

$79 / $129 / $249 · PDF + EPUB · 450 pages

The DynamoDB Book is the book people mean when they say "the DynamoDB book". Its own site describes the Basic package ($79) as "The 450-page book (PDF & EPUB format) with concepts, tips, strategies, and five walkthrough examples" plus "Six printable cheatsheets". The Plus package ($129) adds a 60-page Analytics supplement, a 60-page Operations supplement, and deployable code for each of the five examples; Premium ($249) adds video walkthroughs of every chapter. (Pricing and contents per dynamodbbook.com, verified 2026-08-05.)

The endorsements are unusually strong for a self-published technical book: Rick Houlihan (then Principal Technologist, NoSQL @ AWS), Jeremy Daly, and Corey Quinn all appear on the sales page. AWS's own learning-resources page lists it as "A comprehensive guide from Alex DeBrie that teaches a strategy-driven approach to data modeling with DynamoDB".

What it does best: the five end-to-end walkthroughs. Working a real access-pattern list into a key schema is the skill that separates a DynamoDB table that scales from one that needs a rewrite in eighteen months, and there is very little free material that shows the whole arc. The chapters on single-table design and one-to-many relationships are the reason the book gets recommended.

How old it is: the publicly hosted table of contents is stamped "Version 1.0, 2020-04-06", and as of 2026-10-09 no second edition has been announced. What the books predate below lists the later changes and the four that alter its advice.

Amazon DynamoDB: The Definitive Guide

Packt · August 2024 · 414 pages · ISBN 9781803246895

Amazon DynamoDB: The Definitive Guide by Aman Dhingra and Mike Mackay is the newer of the two, and both authors worked at AWS. The publisher's listing describes Dhingra as a "senior DynamoDB specialist solutions architect at AWS" and Mackay as a "former senior NoSQL specialist solutions architect at AWS" (Leanpub). (Date, page count and ISBN per the O'Reilly listing, verified 2026-08-05.)

It covers modeling (key-value design, moving from a relational mindset, vertical partitioning, secondary indexes) and the operational half: capacity modes, the request routers and storage nodes underneath, backup, streams, global tables, DAX and migrations. The DynamoDB Book is built around five end-to-end walkthroughs (per its site); the Definitive Guide's chapter list puts more of its weight on the operational topics above.

What the books predate

Both books are older than features in current use, and the Definitive Guide's August 2024 date puts it before the November 2024 pricing change that halved on-demand costs.

Sources: the AWS What's New posts for PartiQL, the price reduction, warm throughput and multi-attribute GSI keys, plus the TransactWriteItems API reference, all verified 2026-08-05.

Four of these change the advice: the transaction limit above, and three more. The 50% on-demand price cut moves the break-even point between on-demand and Provisioned capacity capacity, so any capacity-mode guidance written before November 2024 is doing the arithmetic with the old number; see the DynamoDB cost model. Multi-attribute GSI keys let an index key on up to four partition-key and four sort-key attributes, so a GSI no longer needs a hand-concatenated synthetic key for that. And PartiQL gives DynamoDB a SQL-compatible syntax, but it is a query surface, not a relational engine: it still cannot do cross-table JOIN, GROUP BY or aggregates (see PartiQL vs SQL).

Is the 2020 book still worth buying?

Yes, with a caveat.

What has not aged: partition-key design, GSI overloading, item collections, sort-key patterns, one-to-many and many-to-many modeling, and the five walkthroughs. The base table's key model has not changed since the book was written, and modeling is what the book is about. One change touches it: since November 2025 a GSI key can span several attributes, so where the book builds a concatenated sort key for an index, check whether a multi-attribute GSI key now does the job.

What has: the feature inventory, the pricing arithmetic, and any statement of the form "DynamoDB cannot do X". Check the table above before trusting one.

Buy it for the modeling. Do not use it as a current reference for limits, prices, or what the API supports; use AWS's service quotas and pricing pages for those, which are free and updated.

The free path

You do not have to buy anything to learn DynamoDB properly.

• dynamodbguide.com: DeBrie's own free primer, described on its front page as "A Primer on the DynamoDB NoSQL database". It covers the concepts and API surface well and costs nothing. It is older than the book.
• The AWS Developer Guide: the only source in this list that is current by construction. Unbeatable as a reference, poor as a teacher of modeling.
• awesome-dynamodb: a curated link list, useful for finding conference talks. It is updated a few times a year (last merge March 2026).

Where to go next

Reading about modeling only gets you so far. At some point you write an access-pattern list down and turn it into a key schema. The free single-table design planner does that step interactively: list your access patterns, get a PK/SK and GSI plan with the CreateTable JSON, and see where a pattern does not fit. It is a fast way to practice the modeling loop the books teach without setting up an AWS account.

From there, DynamoDB Local lets you run the walkthroughs against a real engine for free, and the guide to DynamoDB GUI clients covers the tools for looking at the items you create. If you want to browse those tables and query them with SQL aggregates and joins, which PartiQL can't give you, download DynoTable: start Local with -sharedDb, then point an Offline profile at it. Its SQL Workbench runs INNER and LEFT joins whose target is keyed on a partition key, on a paid plan.