Chapter 82: Stream API in Java

A Stream<T> is a one-use pipeline over a sequence of values. It is not a replacement for List—think assembly line: data flows through stages (filter, map, …) until a terminal operation (collect, forEach, count) finishes the job and closes the line.

Chapters 83–85 zoom into filter, map, and collect. reduce waits in Chapter 86.

1. Topic title

Stream: describe a computation pipeline; nothing runs until a terminal pulls the chain

2. What it means

Source: list.stream(), Stream.of("a","b"), Stream.generate, Files.lines, …

Intermediate ops: return a new stream—lazy, can fuse internally.

Terminal ops: produce a non-stream result or side effect—eager, triggers computation.

Calling filter alone does nothing visible—no terminal, no work.

3. Why it is used

Declarative list processing—say what transformation chain you want without manual index loops.

Composable stages—small lambdas bolt together; readers scan top to bottom like a recipe.

4. Mental picture

A roller coaster car is the element. filter is a height gate, map repaints the car, collect is the photo booth at the end—you only get the picture when the ride finishes.

5. Java code example

import java.util.List;
 
public class StreamIntroDemo {
    public static void main(String[] args) {
        List<Integer> nums = List.of(1, 2, 3, 4, 5);
        long count =
                nums.stream()
                        .peek(n -> System.out.println("see " + n))
                        .filter(n -> n % 2 == 0)
                        .count();
        System.out.println("even count=" + count);
    }
}

peek logs during the terminal count—proof intermediates stayed quiet until count executed.

6. Explanation of code

stream() does not copy the list into a new data structure up front—it sets up views and spliterators that walk the backing data when needed.

7. Common mistakes

Using a stream twice—after terminal, the stream is spent; call list.stream() again.

.parallelStream() everywhere—debug pain and ordering surprises; default sequential until profiling says otherwise.

8. Best practices

Prefer IntStream / LongStream when primitives dominate to reduce boxing.

Keep side effects in peek only for temporary debugging—production pipelines stay pure in intermediates when possible.

9. Small practice task

Build Stream<String> from Stream.of("a","bb","ccc"), map to lengths, then count how many lengths are >= 2—do it with filter on the int stream after mapToInt (preview mapToInt mentally; full detail in Chapter 84).

Beginner tip

If nothing prints inside your filter, verify you actually called a terminal—filter alone is a planned sieve, not a working one yet.