java stream
stream 是 java 8 的一大亮点，专注于了集合的处理，包括抽取，过滤，转化，聚合，化简等，借助新的 lambda 表达式，极大地简化了流式处理的流程，提高了代码的可读性，同时支持并发和串行两种模式，可以很容易地写出高性能的代码
基本概念
stream 相当于一个高级版本的 iterator，iterator 只有简单的遍历功能，而 stream 在这个基础上提供了抽取，过滤，转化，聚合，化简等丰富的流处理功能
常见的操作主要中间操作和终止操作，一个流处理可以会经过若干个中间操作和一个终止操作，中间操作返回的依然是一个流，终止操作获得一个特定的结果，中间操作是惰性的，在遇到终止操作之前，中间操作只是记录一个操作步骤，不会从流中读取任何数据
创建流
// 从容器中创建
Stream stream1 = List.of(1, 2, 3, 4, 5).stream();
Stream stream2 = Stream.of(1, 2, 3, 4, 5);
// 创建随机数
Stream stream3 = new Random().ints().limit(10).boxed();
Stream stream4 = ThreadLocalRandom.current().ints().limit(10).boxed();
// 文件流
Stream stream5 = new BufferedReader(new FileReader("/tmp/test.txt")).lines();
// IntStream
Stream stream6 = IntStream.range(1, 10).boxed();
// generate
Stream stream7 = Stream.generate(() -> (int) System.currentTimeMillis()).limit(10);
// iterate
Stream stream8 = Stream.iterate(1, x -> x + 1).limit(10);
Stream stream9 = Stream.iterate(1, x -> x < 10, x -> x + 1);
终止操作
终止操作代表一个计算的结果，常见的终止操作如下:
-
forEach: 遍历元素 -
forEachOrdered: 按顺序遍历元素，主要在并发场景下使用 -
toArray: 将流输出到数组中 -
collect: 收集流，可以输出到指定的Collector中，Collectors中自带了toList，toSet等Collector -
anyMatch: 任意元素满足谓词，返回 true -
allMatch: 所有元素满足谓词，返回 true -
noneMatch: 所有元素都不满足谓词，返回 true -
count: 元素数量 -
findFirst: 返回第一个元素 -
findAny: 返回任意一个元素 -
min: 返回最小元素 -
max: 返回最大元素 -
reduce: 聚合，将流中所有的数据聚合成一个值
Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).forEach(System.out::print);
Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).forEachOrdered(System.out::print);
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).toArray(), equalTo(new Integer[]{1, 2, 3, 4, 5, 6, 7, 8, 9}));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).collect(Collectors.toList()), equalTo(List.of(1, 2, 3, 4, 5, 6, 7, 8, 9)));
assertFalse(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).anyMatch(x -> x > 10));
assertTrue(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).allMatch(x -> x < 10));
assertTrue(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).noneMatch(x -> x > 10));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).count(), 9);
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).findFirst().orElse(0), Integer.valueOf(1));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).findAny().orElse(0), Integer.valueOf(1));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).iterator().next(), Integer.valueOf(1));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).max(Comparator.comparingInt(x -> x)).orElse(0), Integer.valueOf(9));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).min(Comparator.comparingInt(x -> x)).orElse(0), Integer.valueOf(1));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).reduce((x, y) -> x + y).orElse(0), Integer.valueOf(45));
中间操作
-
map: 转化，将流中元素按照 Founction 映射成另一种元素 -
filter: 过滤，将不符合条件的元素从流中删除 -
distinct: 去重，去掉重复元素 -
sorted: 排序 -
limit: 最多取元素的个数 -
skip: 跳过前 n 个元素 -
peek: 和 forEach 操作类似，但依然会返回当前流 -
takeWhile: 获取元素直到不满足条件 -
dropWhile: 丢弃元素直到满足条件
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).map(x -> x * x).collect(Collectors.toList()), equalTo(List.of(
1, 4, 9, 16, 25, 36, 49, 64, 81
)));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).filter(x -> x % 2 == 0).collect(Collectors.toList()), equalTo(List.of(
2, 4, 6, 8
)));
assertThat(Stream.of(4, 1, 2, 1, 2, 4, 3, 3).distinct().collect(Collectors.toList()), equalTo(List.of(
4, 1, 2, 3
)));
assertThat(Stream.of(6, 4, 7, 3, 2, 9, 1, 5, 8).sorted().collect(Collectors.toList()), equalTo(List.of(
1, 2, 3, 4, 5, 6, 7, 8, 9
)));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).limit(3).collect(Collectors.toList()), equalTo(List.of(
1, 2, 3
)));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).skip(6).collect(Collectors.toList()), equalTo(List.of(
7, 8, 9
)));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).peek(System.out::println).collect(Collectors.toList()), equalTo(List.of(
1, 2, 3, 4, 5, 6, 7, 8, 9
)));
assertEquals(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).mapToInt(x -> x * x).sum(), 285);
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).takeWhile(x -> x < 5).collect(Collectors.toList()), equalTo(List.of(
1, 2, 3, 4
)));
assertThat(Stream.of(1, 2, 3, 4, 5, 6, 7, 8, 9).dropWhile(x -> x < 5).collect(Collectors.toList()), equalTo(List.of(
5, 6, 7, 8, 9
)));
group by
类似于数据库中 group by 的操作，流处理也支持 group by，能根据对象中的某个字段聚合
class Student {
private final String grade;
private final String name;
private final int chinese;
private final int english;
private Student(String grade, String name, int chinese, int english) {
this.grade = grade;
this.name = name;
this.chinese = chinese;
this.english = english;
}
@Override
public String toString() {
return "[" + grade + " " + name + " " + chinese + " " + english + "]";
}
}
Random random = new Random();
Stream stream = Stream.generate(() ->
new Student(
"grade" + Math.abs(random.nextInt() % 4 + 1),
"student" + Math.abs(random.nextInt() % 1000),
Math.abs(random.nextInt() % 30 + 70),
Math.abs(random.nextInt() % 30 + 70)
)
);
Map> map = stream.limit(10).collect(Collectors.groupingBy(
x -> x.grade, Collectors.toList()
));
map.forEach((k, v) -> {
System.out.println(k + " => " + v);
});
链接
- 测试代码: https://github.com/hatlonely/hellojava/blob/master/src/test/java/util/StreamTest.java