What is Functional Reactive Programming?
- Also known as data-flow programming, builds on concepts of Functional Programming
- Is a combination of Functional Programming and Reactive Programming
- It allows you to react to data changes over time using declarative, functional transformations
Think of it like a spreadsheet:
- If cell A changes, cell B (which depends on A) updates automatically.
- You don’t need to manually update cell B—the relationship is declared and updates reactively.
- FRP works similarly: data flows automatically based on dependencies.
FRP allows us to express data flows over time using declarative, functional transformations.
🔗 Further Reading: Wikipedia - Functional Reactive Programming
Differences Between FRP and FP
| Feature | Functional Programming (FP) | Functional Reactive Programming (FRP) |
|---|---|---|
| Focus | Data transformation | Data flow over time |
| Execution Model | Applies transformations on existing data | Continuously reacts to new data |
| State Management | Stateless, pure functions | Reactive streams manage state changes |
| Mutation | Avoids mutation, but still works on fixed data | Avoids mutation, and data is event-driven |
Example
import Combine
// 1. Define a reactive stream for user input
let userInput = PassthroughSubject<String, Never>()
// 2. Apply functional transformations (functional part)
let validatedInput = userInput
.filter { $0.count > 3 } // Only allow strings longer than 3 characters
.map { $0.uppercased() } // Convert to uppercase
// 3. Subscribe to changes (reactive part)
let cancellable = validatedInput
.sink { print("📢 Valid Input Received: \($0)") }
// 4. Simulate user typing (reactive updates)
userInput.send("Hi") // No output (filtered out)
userInput.send("Hello") // Output: 📢 Valid Input Received: HELLO
userInput.send("Swift") // Output: 📢 Valid Input Received: SWIFT
Why is this FRP?
✅ Functional → Uses filter and map to transform data.
✅ Reactive → The subscriber automatically reacts to new values.
✅ Event-Driven → The program reacts as input changes, instead of executing once.