One Assertion Per Test vs Multiple Assertions
You'll learn: When to use a single assertion per test versus grouping multiple assertions together, and the trade-offs involved.
The Core Question
After you Act (execute the behavior), how many checks should you make in the Assert phase? Should each test verify exactly one thing with one assertion, or can you group related assertions together?
The Single-Assertion Perspective
Some developers advocate for one assertion per test. The reasoning: when a test fails, you know exactly what broke. If a test named testUserEmailIsValidated contains only assert(user.email).isValid(), failure pinpoints the problem immediately. This approach makes tests extremely focused and failure messages crystal-clear.
The Multiple-Assertions Reality
However, strict one-assertion-per-test can be impractical. Imagine testing that a createUser function properly initializes a user object. You'd reasonably want to verify:
assert(user.name).equals("Alice")
assert(user.email).equals("alice@example.com")
assert(user.isActive).isTrue()
These assertions test one behavior (user creation), just from multiple angles. Splitting into three separate tests would mean duplicating setup three times and obscuring that you're verifying a single concept.
The Practical Balance
Test one behavior, not necessarily one assertion. If multiple assertions verify different facets of the same outcome—like checking an object's multiple properties after one action—keep them together. But if assertions verify unrelated behaviors, split them into separate tests.
Think of it like checking a package delivery: confirming the box arrived, contains the right item, and isn't damaged are all part of verifying "successful delivery," not three separate concerns.
Key Takeaway: Prefer one assertion per test for clarity, but group multiple assertions when they collectively verify a single, cohesive behavior—prioritize test readability and avoiding duplication over rigid rules.