Return to site

🗄️⚠️ SPRING DATA findAll(): CONVENIENT, BUT NOT ALWAYS SAFE Inspired by Java Champion Vlad Mihalcea.

July 17, 2026

Spring Data makes repository creation incredibly easy:

But this also exposes findAll() by default.

That method is not automatically bad. The danger appears when it becomes the default answer for querying a table whose size can grow dramatically in production.

🔸 TL;DR

▪️ findAll() is convenient, not universally wrong.

▪️ Fetching an entire large table to filter in memory is the real anti-pattern.

▪️ Query only the rows and columns required by the use case.

▪️ Repository APIs should prevent dangerous shortcuts, not encourage them.

🔸 THE ANTI-PATTERN

The application loads every row, creates entities, may trigger additional lazy-loading queries and then discards most of the data.

The database is optimized for filtering. Java should not replace the WHERE clause.

🔸 PUSH FILTERING TO THE DATABASE

Only the required data crosses the network. A projection also avoids loading complete entities when only titles are needed.

🔸 DESIGN A SAFER REPOSITORY API

Expose only operations that make sense by default. Add findAll() explicitly for bounded reference tables where loading every row is an informed decision.

🔸 TAKEAWAYS

▪️ Think about future production volume, not today’s test dataset.

▪️ Watch for N+1 queries caused by lazy associations.

▪️ Prefer derived queries, JPQL, SQL, Specifications or projections.

▪️ Keep findAll() only where “all” is genuinely small and intentional.

A repository interface is an API. Every method it exposes becomes an invitation to use it. 🚀

#Java #SpringData #SpringBoot #JPA #Hibernate #SQL #Performance #SoftwareArchitecture #BackendDevelopment #VladMihalcea

Go further with Java certification:

Java👇

Spring👇

SpringBook👇

JavaBook👇