Course contentsShow
← Java
Lesson 1432 of 2,87032. Streams API: Intermediate OperationsFree lesson

mapToInt, mapToLong, mapToDouble: Primitive Specializations

Using specialized map variants to convert object streams to primitive streams for better performance.

mapToInt, mapToLong, mapToDouble: Primitive Specializations

You'll learn how to convert object streams to primitive streams for better performance and memory efficiency.

The Problem with Generic Streams

When you use the regular map operation, you always get a Stream<T> of objects. Even when transforming to numbers, you end up with Stream<Integer>, Stream<Long>, or Stream<Double>. This means boxing overhead—every primitive value gets wrapped in an object, consuming extra memory and CPU cycles.

Primitive Stream Specializations

Java provides three specialized mapping operations that convert to primitive streams:

  • mapToInt() returns an IntStream
  • mapToLong() returns a LongStream
  • mapToDouble() returns a DoubleStream

These methods eliminate boxing, use less memory, and unlock primitive-specific operations like sum(), average(), and max().

Basic Examples

List<String> words = List.of("cat", "elephant", "ox");

// Regular map: Stream<Integer> with boxing
Stream<Integer> lengths = words.stream()
    .map(String::length);

// mapToInt: IntStream without boxing
IntStream lengthsInt = words.stream()
    .mapToInt(String::length);
List<Product> products = getProducts();

// Extract prices as DoubleStream
DoubleStream prices = products.stream()
    .mapToDouble(Product::getPrice);

// Now you can use primitive operations
double total = prices.sum();

When to Use Them

Use these specializations whenever you're mapping to numeric values and plan to perform arithmetic operations or terminal operations like sum(), average(), min(), or max(). The performance gain is especially noticeable with large datasets.

Key Takeaway: Use mapToInt, mapToLong, and mapToDouble instead of regular map when transforming to numbers—they eliminate boxing overhead and provide specialized numeric operations unavailable on generic streams.