Java 8 中的 Streams API 详解
Streams 的背景，以及 Java 8 中的使用详解
<!-- Article Body -->
<h2 id="major1" class="ibm-h2">为什么需要 Stream</h2><p>Stream 作为 Java 8 的一大亮点，它与 java.io 包里的 InputStream 和 OutputStream 是完全不同的概念。它也不同于 StAX 对 XML 解析的
Stream，也不是 Amazon Kinesis 对大数据实时处理的 Stream。Java 8 中的 Stream
是对集合（Collection）对象功能的增强，它专注于对集合对象进行各种非常便利、高效的聚合操作（aggregate operation），或者大批量数据操作 (bulk data
operation)。Stream API 借助于同样新出现的 Lambda
表达式，极大的提高编程效率和程序可读性。同时它提供串行和并行两种模式进行汇聚操作，并发模式能够充分利用多核处理器的优势，使用 fork/join
并行方式来拆分任务和加速处理过程。通常编写并行代码很难而且容易出错, 但使用 Stream API 无需编写一行多线程的代码，就可以很方便地写出高性能的并发程序。所以说，Java 8
中首次出现的 java.util.stream 是一个函数式语言+多核时代综合影响的产物。</p><h3 id="major2" class="ibm-h3">什么是聚合操作</h3><p>在传统的 J2EE 应用中，Java 代码经常不得不依赖于关系型数据库的聚合操作来完成诸如：</p><ul class="ibm-bullet-list"><li> 客户每月平均消费金额</li><li> 最昂贵的在售商品</li><li> 本周完成的有效订单（排除了无效的）</li><li> 取十个数据样本作为首页推荐</li></ul><p>这类的操作。</p><p>但在当今这个数据大爆炸的时代，在数据来源多样化、数据海量化的今天，很多时候不得不脱离 RDBMS，或者以底层返回的数据为基础进行更上层的数据统计。而 Java 的集合 API
中，仅仅有极少量的辅助型方法，更多的时候是程序员需要用 Iterator 来遍历集合，完成相关的聚合应用逻辑。这是一种远不够高效、笨拙的方法。在 Java 7 中，如果要发现 type 为
grocery 的所有交易，然后返回以交易值降序排序好的交易 ID 集合，我们需要这样写：</p><h5 id="listing1" class="ibm-h5">清单 1. Java 7 的排序、取值实现</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_394359" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div><div class="line number12 index11 alt1">12</div><div class="line number13 index12 alt2">13</div><div class="line number14 index13 alt1">14</div><div class="line number15 index14 alt2">15</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Transaction</code><code class="htmlscript plain">> groceryTransactions = new Arraylist<>();</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">for(Transaction t: transactions){</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">if(t.getType() == Transaction.GROCERY){</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">groceryTransactions.add(t);</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">}</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">Collections.sort(groceryTransactions, new Comparator(){</code></div><div class="line number8 index7 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">public int compare(Transaction t1, Transaction t2){</code></div><div class="line number9 index8 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">return t2.getValue().compareTo(t1.getValue());</code></div><div class="line number10 index9 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number11 index10 alt2"><code class="htmlscript plain">});</code></div><div class="line number12 index11 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> transactionIds = new ArrayList<>();</code></div><div class="line number13 index12 alt2"><code class="htmlscript plain">for(Transaction t: groceryTransactions){</code></div><div class="line number14 index13 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">transactionsIds.add(t.getId());</code></div><div class="line number15 index14 alt2"><code class="htmlscript plain">}</code></div></div></td></tr></tbody></table></div></div></div></span><p>而在 Java 8 使用 Stream，代码更加简洁易读；而且使用并发模式，程序执行速度更快。</p><h5 id="listing2" class="ibm-h5">清单 2. Java 8 的排序、取值实现</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_911815" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> transactionsIds = transactions.parallelStream().</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">filter(t -> t.getType() == Transaction.GROCERY).</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">sorted(comparing(Transaction::getValue).reversed()).</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">map(Transaction::getId).</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">collect(toList());</code></div></div></td></tr></tbody></table></div></div></div></span><h2 id="major3" class="ibm-h2">Stream 总览</h2><h3 id="major4" class="ibm-h3">什么是流</h3><p>Stream 不是集合元素，它不是数据结构并不保存数据，它是有关算法和计算的，它更像一个高级版本的 Iterator。原始版本的
Iterator，用户只能显式地一个一个遍历元素并对其执行某些操作；高级版本的 Stream，用户只要给出需要对其包含的元素执行什么操作，比如 “过滤掉长度大于 10
的字符串”、“获取每个字符串的首字母”等，Stream 会隐式地在内部进行遍历，做出相应的数据转换。</p><p>Stream 就如同一个迭代器（Iterator），单向，不可往复，数据只能遍历一次，遍历过一次后即用尽了，就好比流水从面前流过，一去不复返。</p><p>而和迭代器又不同的是，Stream 可以并行化操作，迭代器只能命令式地、串行化操作。顾名思义，当使用串行方式去遍历时，每个 item 读完后再读下一个
item。而使用并行去遍历时，数据会被分成多个段，其中每一个都在不同的线程中处理，然后将结果一起输出。Stream 的并行操作依赖于 Java7 中引入的 Fork/Join
框架（JSR166y）来拆分任务和加速处理过程。Java 的并行 API 演变历程基本如下：</p><ol type="1"><li> 1.0-1.4 中的 java.lang.Thread</li><li> 5.0 中的 java.util.concurrent</li><li> 6.0 中的 Phasers 等</li><li> 7.0 中的 Fork/Join 框架</li><li> 8.0 中的 Lambda</li></ol><p>Stream 的另外一大特点是，数据源本身可以是无限的。</p><h3 id="major5" class="ibm-h3">流的构成</h3><p>当我们使用一个流的时候，通常包括三个基本步骤：</p><p>获取一个数据源（source）→ 数据转换→执行操作获取想要的结果，每次转换原有 Stream 对象不改变，返回一个新的 Stream
对象（可以有多次转换），这就允许对其操作可以像链条一样排列，变成一个管道，如下图所示。</p><h5 id="img001" class="ibm-h5">图 1. 流管道 (Stream Pipeline) 的构成</h5><img src="https://images2018.cnblogs.com/blog/1112483/201807/1112483-20180720131435137-108803321.png" class="ibm-downsize" alt="图 1. 流管道 (Stream Pipeline) 的构成" height="284" width="331"><p class="ibm-ind-link ibm-hide"><a class="ibm-popup-link" onclick="IBMCore.common.widget.overlay.show('N100AF');return false;" href="#N100AF">点击查看大图</a></p><p>有多种方式生成 Stream Source：</p><ul class="ibm-bullet-list"><li> 从 Collection 和数组</li></ul><ul class="ibm-bullet-list"><li><ul class="ibm-bullet-list"><li> Collection.stream()</li><li> Collection.parallelStream()</li><li> Arrays.stream(T array) or Stream.of()</li></ul> 从 BufferedReader<ul class="ibm-bullet-list"><li> java.io.BufferedReader.lines()</li></ul></li></ul><ul class="ibm-bullet-list"><li> 静态工厂</li></ul><ul class="ibm-bullet-list"><li> java.util.stream.IntStream.range()</li><li> java.nio.file.Files.walk()</li></ul><ul class="ibm-bullet-list"><li> 自己构建</li></ul><ul class="ibm-bullet-list"><li><ul class="ibm-bullet-list"><li> java.util.Spliterator</li></ul> 其它<ul class="ibm-bullet-list"><li> Random.ints()</li><li> BitSet.stream()</li><li> Pattern.splitAsStream(java.lang.CharSequence)</li><li> JarFile.stream()</li></ul></li></ul><p>流的操作类型分为两种：</p><ul class="ibm-bullet-list"><li><strong>Intermediate</strong>：一个流可以后面跟随零个或多个 intermediate
操作。其目的主要是打开流，做出某种程度的数据映射/过滤，然后返回一个新的流，交给下一个操作使用。这类操作都是惰性化的（lazy），就是说，仅仅调用到这类方法，并没有真正开始流的遍历。</li><li><strong>Terminal</strong>：一个流只能有一个 terminal
操作，当这个操作执行后，流就被使用“光”了，无法再被操作。所以这必定是流的最后一个操作。Terminal 操作的执行，才会真正开始流的遍历，并且会生成一个结果，或者一个 side
effect。</li></ul><p>在对于一个 Stream 进行多次转换操作 (Intermediate 操作)，每次都对 Stream 的每个元素进行转换，而且是执行多次，这样时间复杂度就是 N（转换次数）个 for
循环里把所有操作都做掉的总和吗？其实不是这样的，转换操作都是 lazy 的，多个转换操作只会在 Terminal 操作的时候融合起来，一次循环完成。我们可以这样简单的理解，Stream
里有个操作函数的集合，每次转换操作就是把转换函数放入这个集合中，在 Terminal 操作的时候循环 Stream 对应的集合，然后对每个元素执行所有的函数。</p><p>还有一种操作被称为 <strong>short-circuiting</strong>。用以指：</p><ul class="ibm-bullet-list"><li> 对于一个 intermediate 操作，如果它接受的是一个无限大（infinite/unbounded）的 Stream，但返回一个有限的新 Stream。</li><li> 对于一个 terminal 操作，如果它接受的是一个无限大的 Stream，但能在有限的时间计算出结果。</li></ul><p>当操作一个无限大的 Stream，而又希望在有限时间内完成操作，则在管道内拥有一个 short-circuiting 操作是必要非充分条件。</p><h5 id="listing3" class="ibm-h5">清单 3. 一个流操作的示例</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_929573" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">int sum = widgets.stream()</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">.filter(w -> w.getColor() == RED)</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.mapToInt(w -> w.getWeight())</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.sum();</code></div></div></td></tr></tbody></table></div></div></div></span><p>stream() 获取当前小物件的 source，filter 和 mapToInt 为 intermediate 操作，进行数据筛选和转换，最后一个 sum() 为 terminal
操作，对符合条件的全部小物件作重量求和。</p><h2 id="major6" class="ibm-h2">流的使用详解</h2><p>简单说，对 Stream 的使用就是实现一个 filter-map-reduce 过程，产生一个最终结果，或者导致一个副作用（side effect）。</p><h3 id="major7" class="ibm-h3">流的构造与转换</h3><p>下面提供最常见的几种构造 Stream 的样例。</p><h5 id="listing4" class="ibm-h5">清单 4. 构造流的几种常见方法</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_237386" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">// 1. Individual values</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Stream stream = Stream.of("a", "b", "c");</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">// 2. Arrays</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">String [] strArray = new String[] {"a", "b", "c"};</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">stream = Stream.of(strArray);</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">stream = Arrays.stream(strArray);</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">// 3. Collections</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> list = Arrays.asList(strArray);</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">stream = list.stream();</code></div></div></td></tr></tbody></table></div></div></div></span><p>需要注意的是，对于基本数值型，目前有三种对应的包装类型 Stream：</p><p>IntStream、LongStream、DoubleStream。当然我们也可以用 Stream<Integer>、Stream<Long>
>、Stream<Double>，但是 boxing 和 unboxing 会很耗时，所以特别为这三种基本数值型提供了对应的 Stream。</p><p>Java 8 中还没有提供其它数值型 Stream，因为这将导致扩增的内容较多。而常规的数值型聚合运算可以通过上面三种 Stream 进行。</p><h5 id="listing5" class="ibm-h5">清单 5. 数值流的构造</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_438603" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">IntStream.of(new int[]{1, 2, 3}).forEach(System.out::println);</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">IntStream.range(1, 3).forEach(System.out::println);</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">IntStream.rangeClosed(1, 3).forEach(System.out::println);</code></div></div></td></tr></tbody></table></div></div></div></span><h5 id="listing6" class="ibm-h5">清单 6. 流转换为其它数据结构</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_696605" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">// 1. Array</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">String[] strArray1 = stream.toArray(String[]::new);</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">// 2. Collection</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> list1 = stream.collect(Collectors.toList());</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> list2 = stream.collect(Collectors.toCollection(ArrayList::new));</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">Set set1 = stream.collect(Collectors.toSet());</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">Stack stack1 = stream.collect(Collectors.toCollection(Stack::new));</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">// 3. String</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">String str = stream.collect(Collectors.joining()).toString();</code></div></div></td></tr></tbody></table></div></div></div></span><p>一个 Stream 只可以使用一次，上面的代码为了简洁而重复使用了数次。</p><h3 id="major8" class="ibm-h3">流的操作</h3><p>接下来，当把一个数据结构包装成 Stream 后，就要开始对里面的元素进行各类操作了。常见的操作可以归类如下。</p><ul class="ibm-bullet-list"><li> Intermediate：</li></ul><p>map (mapToInt, flatMap 等)、 filter、 distinct、 sorted、 peek、 limit、 skip、 parallel、
sequential、 unordered</p><ul class="ibm-bullet-list"><li> Terminal：</li></ul><p>forEach、 forEachOrdered、 toArray、 reduce、 collect、 min、 max、 count、 anyMatch、
allMatch、 noneMatch、 findFirst、 findAny、 iterator</p><ul class="ibm-bullet-list"><li> Short-circuiting：</li></ul><p>anyMatch、 allMatch、 noneMatch、 findFirst、 findAny、 limit</p><p>我们下面看一下 Stream 的比较典型用法。</p><p><strong>map/flatMap</strong></p><p>我们先来看 map。如果你熟悉 scala 这类函数式语言，对这个方法应该很了解，它的作用就是把 input Stream 的每一个元素，映射成 output Stream
的另外一个元素。</p><h5 id="listing7" class="ibm-h5">清单 7. 转换大写</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_595455" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> output = wordList.stream().</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">map(String::toUpperCase).</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">collect(Collectors.toList());</code></div></div></td></tr></tbody></table></div></div></div></span><p>这段代码把所有的单词转换为大写。</p><h5 id="listing8" class="ibm-h5">清单 8. 平方数</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_928732" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> nums = Arrays.asList(1, 2, 3, 4);</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> squareNums = nums.stream().</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">map(n -> n * n).</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">collect(Collectors.toList());</code></div></div></td></tr></tbody></table></div></div></div></span><p>这段代码生成一个整数 list 的平方数 {1, 4, 9, 16}。</p><p>从上面例子可以看出，map 生成的是个 1:1 映射，每个输入元素，都按照规则转换成为另外一个元素。还有一些场景，是一对多映射关系的，这时需要 flatMap。</p><h5 id="listing9" class="ibm-h5">清单 9. 一对多</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_992977" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Stream<</code><code class="htmlscript plain">List</code><code class="htmlscript plain"><Integer>> inputStream = Stream.of(</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Arrays.asList(1),</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Arrays.asList(2, 3),</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Arrays.asList(4, 5, 6)</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">);</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">Stream<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> outputStream = inputStream.</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">flatMap((childList) -> childList.stream());</code></div></div></td></tr></tbody></table></div></div></div></span><p>flatMap 把 input Stream 中的层级结构扁平化，就是将最底层元素抽出来放到一起，最终 output 的新 Stream 里面已经没有 List
了，都是直接的数字。</p><p><strong>filter</strong></p><p>filter 对原始 Stream 进行某项测试，通过测试的元素被留下来生成一个新 Stream。</p><h5 id="listing10" class="ibm-h5">清单 10. 留下偶数</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_248884" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Integer[] sixNums = {1, 2, 3, 4, 5, 6};</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Integer[] evens =</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">Stream.of(sixNums).filter(n -> n%2 == 0).toArray(Integer[]::new);</code></div></div></td></tr></tbody></table></div></div></div></span><p>经过条件“被 2 整除”的 filter，剩下的数字为 {2, 4, 6}。</p><h5 id="listing11" class="ibm-h5">清单 11. 把单词挑出来</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_794772" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> output = reader.lines().</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">flatMap(line -> Stream.of(line.split(REGEXP))).</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">filter(word -> word.length() > 0).</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">collect(Collectors.toList());</code></div></div></td></tr></tbody></table></div></div></div></span><p>这段代码首先把每行的单词用 flatMap 整理到新的 Stream，然后保留长度不为 0 的，就是整篇文章中的全部单词了。</p><p><strong>forEach</strong></p><p>forEach 方法接收一个 Lambda 表达式，然后在 Stream 的每一个元素上执行该表达式。</p><h5 id="listing12" class="ibm-h5">清单 12. 打印姓名（forEach 和 pre-java8 的对比）</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_220482" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">// Java 8</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">roster.stream()</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.filter(p -> p.getGender() == Person.Sex.MALE)</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.forEach(p -> System.out.println(p.getName()));</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">// Pre-Java 8</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">for (Person p : roster) {</code></div><div class="line number7 index6 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">if (p.getGender() == Person.Sex.MALE) {</code></div><div class="line number8 index7 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">System.out.println(p.getName());</code></div><div class="line number9 index8 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">}</code></div></div></td></tr></tbody></table></div></div></div></span><p>对一个人员集合遍历，找出男性并打印姓名。可以看出来，forEach 是为 Lambda 而设计的，保持了最紧凑的风格。而且 Lambda
表达式本身是可以重用的，非常方便。当需要为多核系统优化时，可以 parallelStream().forEach()，只是此时原有元素的次序没法保证，并行的情况下将改变串行时操作的行为，此时
forEach 本身的实现不需要调整，而 Java8 以前的 for 循环 code 可能需要加入额外的多线程逻辑。</p><p>但一般认为，forEach 和常规 for 循环的差异不涉及到性能，它们仅仅是函数式风格与传统 Java 风格的差别。</p><p>另外一点需要注意，forEach 是 terminal 操作，因此它执行后，Stream 的元素就被“消费”掉了，你无法对一个 Stream 进行两次 terminal
运算。下面的代码是错误的：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_342995" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">stream.forEach(element -> doOneThing(element));</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">stream.forEach(element -> doAnotherThing(element));</code></div></div></td></tr></tbody></table></div></div></div></span><p>相反，具有相似功能的 intermediate 操作 peek 可以达到上述目的。如下是出现在该 api javadoc 上的一个示例。</p><h5 id="listing13" class="ibm-h5">清单 13. peek 对每个元素执行操作并返回一个新的 Stream</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_596572" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Stream.of("one", "two", "three", "four")</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.filter(e -> e.length() > 3)</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.peek(e -> System.out.println("Filtered value: " + e))</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.map(String::toUpperCase)</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.peek(e -> System.out.println("Mapped value: " + e))</code></div><div class="line number6 index5 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">.collect(Collectors.toList());</code></div></div></td></tr></tbody></table></div></div></div></span><p>forEach 不能修改自己包含的本地变量值，也不能用 break/return 之类的关键字提前结束循环。</p><p><strong>findFirst</strong></p><p>这是一个 termimal 兼 short-circuiting 操作，它总是返回 Stream 的第一个元素，或者空。</p><p>这里比较重点的是它的返回值类型：Optional。这也是一个模仿 Scala 语言中的概念，作为一个容器，它可能含有某值，或者不包含。使用它的目的是尽可能避免
NullPointerException。</p><h5 id="listing14" class="ibm-h5">清单 14. Optional 的两个用例</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_328816" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div><div class="line number12 index11 alt1">12</div><div class="line number13 index12 alt2">13</div><div class="line number14 index13 alt1">14</div><div class="line number15 index14 alt2">15</div><div class="line number16 index15 alt1">16</div><div class="line number17 index16 alt2">17</div><div class="line number18 index17 alt1">18</div><div class="line number19 index18 alt2">19</div><div class="line number20 index19 alt1">20</div><div class="line number21 index20 alt2">21</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">String strA = " abcd ", strB = null;</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">print(strA);</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">print("");</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">print(strB);</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">getLength(strA);</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">getLength("");</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">getLength(strB);</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">public static void print(String text) {</code></div><div class="line number9 index8 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">// Java 8</code></div><div class="line number10 index9 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Optional.ofNullable(text).ifPresent(System.out::println);</code></div><div class="line number11 index10 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">// Pre-Java 8</code></div><div class="line number12 index11 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">if (text != null) {</code></div><div class="line number13 index12 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">System.out.println(text);</code></div><div class="line number14 index13 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number15 index14 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number16 index15 alt1"><code class="htmlscript plain">public static int getLength(String text) {</code></div><div class="line number17 index16 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">// Java 8</code></div><div class="line number18 index17 alt1"><code class="htmlscript plain">return Optional.ofNullable(text).map(String::length).orElse(-1);</code></div><div class="line number19 index18 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">// Pre-Java 8</code></div><div class="line number20 index19 alt1"><code class="htmlscript plain">// return if (text != null) ? text.length() : -1;</code></div><div class="line number21 index20 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">};</code></div></div></td></tr></tbody></table></div></div></div></span><p>在更复杂的 if (xx != null) 的情况中，使用 Optional 代码的可读性更好，而且它提供的是编译时检查，能极大的降低 NPE 这种 Runtime Exception
对程序的影响，或者迫使程序员更早的在编码阶段处理空值问题，而不是留到运行时再发现和调试。</p><p>Stream 中的 findAny、max/min、reduce 等方法等返回 Optional 值。还有例如 IntStream.average() 返回
OptionalDouble 等等。</p><p><strong>reduce</strong></p><p>这个方法的主要作用是把 Stream 元素组合起来。它提供一个起始值（种子），然后依照运算规则（BinaryOperator），和前面 Stream 的第一个、第二个、第 n
个元素组合。从这个意义上说，字符串拼接、数值的 sum、min、max、average 都是特殊的 reduce。例如 Stream 的 sum 就相当于</p><p>Integer sum = integers.reduce(0, (a, b) -> a+b); 或</p><p>Integer sum = integers.reduce(0, Integer::sum);</p><p>也有没有起始值的情况，这时会把 Stream 的前面两个元素组合起来，返回的是 Optional。</p><h5 id="listing15" class="ibm-h5">清单 15. reduce 的用例</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_938096" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div><div class="line number12 index11 alt1">12</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">// 字符串连接，concat = "ABCD"</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">String concat = Stream.of("A", "B", "C", "D").reduce("", String::concat); </code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">// 求最小值，minValue = -3.0</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">double minValue = Stream.of(-1.5, 1.0, -3.0, -2.0).reduce(Double.MAX_VALUE, Double::min); </code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">// 求和，sumValue = 10, 有起始值</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">int sumValue = Stream.of(1, 2, 3, 4).reduce(0, Integer::sum);</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">// 求和，sumValue = 10, 无起始值</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">sumValue = Stream.of(1, 2, 3, 4).reduce(Integer::sum).get();</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">// 过滤，字符串连接，concat = "ace"</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">concat = Stream.of("a", "B", "c", "D", "e", "F").</code></div><div class="line number11 index10 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">filter(x -> x.compareTo("Z") > 0).</code></div><div class="line number12 index11 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">reduce("", String::concat);</code></div></div></td></tr></tbody></table></div></div></div></span><p>上面代码例如第一个示例的 reduce()，第一个参数（空白字符）即为起始值，第二个参数（String::concat）为 BinaryOperator。这类有起始值的 reduce()
都返回具体的对象。而对于第四个示例没有起始值的 reduce()，由于可能没有足够的元素，返回的是 Optional，请留意这个区别。</p><p><strong>limit/skip</strong></p><p>limit 返回 Stream 的前面 n 个元素；skip 则是扔掉前 n 个元素（它是由一个叫 subStream 的方法改名而来）。</p><h5 id="listing16" class="ibm-h5">清单 16. limit 和 skip 对运行次数的影响</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_636518" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div><div class="line number12 index11 alt1">12</div><div class="line number13 index12 alt2">13</div><div class="line number14 index13 alt1">14</div><div class="line number15 index14 alt2">15</div><div class="line number16 index15 alt1">16</div><div class="line number17 index16 alt2">17</div><div class="line number18 index17 alt1">18</div><div class="line number19 index18 alt2">19</div><div class="line number20 index19 alt1">20</div><div class="line number21 index20 alt2">21</div><div class="line number22 index21 alt1">22</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">public void testLimitAndSkip() {</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> persons = new ArrayList();</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">for (int i = 1; i <= 10000; i++) {</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Person person = new Person(i, "name" + i);</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">persons.add(person);</code></div><div class="line number6 index5 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> personList2 = persons.stream().</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">map(Person::getName).limit(10).skip(3).collect(Collectors.toList());</code></div><div class="line number9 index8 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">System.out.println(personList2);</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">}</code></div><div class="line number11 index10 alt2"><code class="htmlscript plain">private class Person {</code></div><div class="line number12 index11 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">public int no;</code></div><div class="line number13 index12 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">private String name;</code></div><div class="line number14 index13 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">public Person (int no, String name) {</code></div><div class="line number15 index14 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">this.no = no;</code></div><div class="line number16 index15 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">this.name = name;</code></div><div class="line number17 index16 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number18 index17 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">public String getName() {</code></div><div class="line number19 index18 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">System.out.println(name);</code></div><div class="line number20 index19 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">return name;</code></div><div class="line number21 index20 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number22 index21 alt1"><code class="htmlscript plain">}</code></div></div></td></tr></tbody></table></div></div></div></span><p>输出结果为：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_371363" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">name1</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">name2</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">name3</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">name4</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">name5</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">name6</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">name7</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">name8</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">name9</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">name10</code></div><div class="line number11 index10 alt2"><code class="htmlscript plain">[name4, name5, name6, name7, name8, name9, name10]</code></div></div></td></tr></tbody></table></div></div></div></span><p>这是一个有 10，000 个元素的 Stream，但在 short-circuiting 操作 limit 和 skip 的作用下，管道中 map 操作指定的 getName()
方法的执行次数为 limit 所限定的 10 次，而最终返回结果在跳过前 3 个元素后只有后面 7 个返回。</p><p>有一种情况是 limit/skip 无法达到 short-circuiting 目的的，就是把它们放在 Stream 的排序操作后，原因跟 sorted 这个 intermediate
操作有关：此时系统并不知道 Stream 排序后的次序如何，所以 sorted 中的操作看上去就像完全没有被 limit 或者 skip 一样。</p><h5 id="listing17" class="ibm-h5">清单 17. limit 和 skip 对 sorted 后的运行次数无影响</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_296827" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> persons = new ArrayList();</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">for (int i = 1; i <= 5; i++) {</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Person person = new Person(i, "name" + i);</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">persons.add(person);</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> personList2 = persons.stream().sorted((p1, p2) -> </code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">p1.getName().compareTo(p2.getName())).limit(2).collect(Collectors.toList());</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">System.out.println(personList2);</code></div></div></td></tr></tbody></table></div></div></div></span><p>上面的示例对清单 13 做了微调，首先对 5 个元素的 Stream 排序，然后进行 limit 操作。输出结果为：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_382327" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">name2</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">name1</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">name3</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">name2</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">name4</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">name3</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">name5</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">name4</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">[stream.StreamDW$Person@816f27d, stream.StreamDW$Person@87aac27]</code></div></div></td></tr></tbody></table></div></div></div></span><p>即虽然最后的返回元素数量是 2，但整个管道中的 sorted 表达式执行次数没有像前面例子相应减少。</p><p>最后有一点需要注意的是，对一个 parallel 的 Steam 管道来说，如果其元素是有序的，那么 limit 操作的成本会比较大，因为它的返回对象必须是前 n
个也有一样次序的元素。取而代之的策略是取消元素间的次序，或者不要用 parallel Stream。</p><p><strong>sorted</strong></p><p>对 Stream 的排序通过 sorted 进行，它比数组的排序更强之处在于你可以首先对 Stream 进行各类 map、filter、limit、skip 甚至 distinct
来减少元素数量后，再排序，这能帮助程序明显缩短执行时间。我们对清单 14 进行优化：</p><h5 id="listing18" class="ibm-h5">清单 18. 优化：排序前进行 limit 和 skip</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_938866" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> persons = new ArrayList();</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">for (int i = 1; i <= 5; i++) {</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Person person = new Person(i, "name" + i);</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">persons.add(person);</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> personList2 = persons.stream().limit(2).sorted((p1, p2) -> p1.getName().compareTo(p2.getName())).collect(Collectors.toList());</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">System.out.println(personList2);</code></div></div></td></tr></tbody></table></div></div></div></span><p>结果会简单很多：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_79758" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">name2</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">name1</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">[stream.StreamDW$Person@6ce253f1, stream.StreamDW$Person@53d8d10a]</code></div></div></td></tr></tbody></table></div></div></div></span><p>当然，这种优化是有 business logic 上的局限性的：即不要求排序后再取值。</p><p><strong>min/max/distinct</strong></p><p>min 和 max 的功能也可以通过对 Stream 元素先排序，再 findFirst 来实现，但前者的性能会更好，为 O(n)，而 sorted 的成本是 O(n log
n)。同时它们作为特殊的 reduce 方法被独立出来也是因为求最大最小值是很常见的操作。</p><h5 id="listing19" class="ibm-h5">清单 19. 找出最长一行的长度</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_343795" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">BufferedReader br = new BufferedReader(new FileReader("c:\SUService.log"));</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">int longest = br.lines().</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">mapToInt(String::length).</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">max().</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">getAsInt();</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">br.close();</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">System.out.println(longest);</code></div></div></td></tr></tbody></table></div></div></div></span><p>下面的例子则使用 distinct 来找出不重复的单词。</p><h5 id="listing20" class="ibm-h5">清单 20. 找出全文的单词，转小写，并排序</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_615558" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">String</code><code class="htmlscript plain">> words = br.lines().</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">flatMap(line -> Stream.of(line.split(" "))).</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">filter(word -> word.length() > 0).</code></div><div class="line number4 index3 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">map(String::toLowerCase).</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">distinct().</code></div><div class="line number6 index5 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">sorted().</code></div><div class="line number7 index6 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">collect(Collectors.toList());</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">br.close();</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">System.out.println(words);</code></div></div></td></tr></tbody></table></div></div></div></span><p><strong>Match</strong></p><p>Stream 有三个 match 方法，从语义上说： </p><ul class="ibm-bullet-list"><li> allMatch：Stream 中全部元素符合传入的 predicate，返回 true</li><li> anyMatch：Stream 中只要有一个元素符合传入的 predicate，返回 true</li><li> noneMatch：Stream 中没有一个元素符合传入的 predicate，返回 true</li></ul><p>它们都不是要遍历全部元素才能返回结果。例如 allMatch 只要一个元素不满足条件，就 skip 剩下的所有元素，返回 false。对清单 13 中的 Person 类稍做修改，加入一个
age 属性和 getAge 方法。</p><h5 id="listing21" class="ibm-h5">清单 21. 使用 Match</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_418183" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div><div class="line number12 index11 alt1">12</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">List<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> persons = new ArrayList();</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">persons.add(new Person(1, "name" + 1, 10));</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">persons.add(new Person(2, "name" + 2, 21));</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">persons.add(new Person(3, "name" + 3, 34));</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">persons.add(new Person(4, "name" + 4, 6));</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">persons.add(new Person(5, "name" + 5, 55));</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">boolean isAllAdult = persons.stream().</code></div><div class="line number8 index7 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">allMatch(p -> p.getAge() > 18);</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">System.out.println("All are adult? " + isAllAdult);</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">boolean isThereAnyChild = persons.stream().</code></div><div class="line number11 index10 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">anyMatch(p -> p.getAge() < 12);</code></div><div class="line number12 index11 alt1"><code class="htmlscript plain">System.out.println("Any child? " + isThereAnyChild);</code></div></div></td></tr></tbody></table></div></div></div></span><p>输出结果：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_410066" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">All are adult? false</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Any child? true</code></div></div></td></tr></tbody></table></div></div></div></span><h3 id="major9" class="ibm-h3">进阶：自己生成流</h3><p><strong>Stream.generate</strong></p><p>通过实现 Supplier 接口，你可以自己来控制流的生成。这种情形通常用于随机数、常量的 Stream，或者需要前后元素间维持着某种状态信息的 Stream。把 Supplier
实例传递给 Stream.generate() 生成的 Stream，默认是串行（相对 parallel 而言）但无序的（相对 ordered 而言）。由于它是无限的，在管道中，必须利用
limit 之类的操作限制 Stream 大小。</p><h5 id="listing22" class="ibm-h5">清单 22. 生成 10 个随机整数</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_292054" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Random seed = new Random();</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Supplier<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">> random = seed::nextInt;</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">Stream.generate(random).limit(10).forEach(System.out::println);</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">//Another way</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">IntStream.generate(() -> (int) (System.nanoTime() % 100)).</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">limit(10).forEach(System.out::println);</code></div></div></td></tr></tbody></table></div></div></div></span><p>Stream.generate() 还接受自己实现的 Supplier。例如在构造海量测试数据的时候，用某种自动的规则给每一个变量赋值；或者依据公式计算 Stream
的每个元素值。这些都是维持状态信息的情形。</p><h5 id="listing23" class="ibm-h5">清单 23. 自实现 Supplier</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_740665" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div><div class="line number11 index10 alt2">11</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Stream.generate(new PersonSupplier()).</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">limit(10).</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">forEach(p -> System.out.println(p.getName() + ", " + p.getAge()));</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">private class PersonSupplier implements Supplier<</code><code class="htmlscript plain">Person</code><code class="htmlscript plain">> {</code></div><div class="line number5 index4 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">private int index = 0;</code></div><div class="line number6 index5 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">private Random random = new Random();</code></div><div class="line number7 index6 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">@Override</code></div><div class="line number8 index7 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">public Person get() {</code></div><div class="line number9 index8 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">return new Person(index++, "StormTestUser" + index, random.nextInt(100));</code></div><div class="line number10 index9 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">}</code></div><div class="line number11 index10 alt2"><code class="htmlscript plain">}</code></div></div></td></tr></tbody></table></div></div></div></span><p>输出结果：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_377561" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div><div class="line number9 index8 alt2">9</div><div class="line number10 index9 alt1">10</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">StormTestUser1, 9</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">StormTestUser2, 12</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">StormTestUser3, 88</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">StormTestUser4, 51</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">StormTestUser5, 22</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">StormTestUser6, 28</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">StormTestUser7, 81</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">StormTestUser8, 51</code></div><div class="line number9 index8 alt2"><code class="htmlscript plain">StormTestUser9, 4</code></div><div class="line number10 index9 alt1"><code class="htmlscript plain">StormTestUser10, 76</code></div></div></td></tr></tbody></table></div></div></div></span><p><strong>Stream.iterate</strong></p><p>iterate 跟 reduce 操作很像，接受一个种子值，和一个 UnaryOperator（例如 f）。然后种子值成为 Stream 的第一个元素，f(seed)
为第二个，f(f(seed)) 第三个，以此类推。</p><p><strong>清单 24. 生成一个等差数列</strong></p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_539081" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Stream.iterate(0, n -> n + 3).limit(10). forEach(x -> System.out.print(x + " "));.</code></div></div></td></tr></tbody></table></div></div></div></span><p>输出结果：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_796840" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">0 3 6 9 12 15 18 21 24 27</code></div></div></td></tr></tbody></table></div></div></div></span><p>与 Stream.generate 相仿，在 iterate 时候管道必须有 limit 这样的操作来限制 Stream 大小。</p><h3 id="major10" class="ibm-h3">进阶：用 Collectors 来进行 reduction 操作</h3><p>java.util.stream.Collectors 类的主要作用就是辅助进行各类有用的 reduction 操作，例如转变输出为 Collection，把 Stream
元素进行归组。</p><p><strong>groupingBy/partitioningBy</strong></p><h5 id="listing25" class="ibm-h5">清单 25. 按照年龄归组</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_855620" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div><div class="line number8 index7 alt1">8</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Map<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">, List<Person>> personGroups = Stream.generate(new PersonSupplier()).</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">limit(100).</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">collect(Collectors.groupingBy(Person::getAge));</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">Iterator it = personGroups.entrySet().iterator();</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">while (it.hasNext()) {</code></div><div class="line number6 index5 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">Map.Entry<</code><code class="htmlscript plain">Integer</code><code class="htmlscript plain">, List<Person>> persons = (Map.Entry) it.next();</code></div><div class="line number7 index6 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">System.out.println("Age " + persons.getKey() + " = " + persons.getValue().size());</code></div><div class="line number8 index7 alt1"><code class="htmlscript plain">}</code></div></div></td></tr></tbody></table></div></div></div></span><p>上面的 code，首先生成 100 人的信息，然后按照年龄归组，相同年龄的人放到同一个 list 中，可以看到如下的输出：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_229447" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div><div class="line number6 index5 alt1">6</div><div class="line number7 index6 alt2">7</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Age 0 = 2</code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Age 1 = 2</code></div><div class="line number3 index2 alt2"><code class="htmlscript plain">Age 5 = 2</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">Age 8 = 1</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">Age 9 = 1</code></div><div class="line number6 index5 alt1"><code class="htmlscript plain">Age 11 = 2</code></div><div class="line number7 index6 alt2"><code class="htmlscript plain">……</code></div></div></td></tr></tbody></table></div></div></div></span><h5 id="listing26" class="ibm-h5">清单 26. 按照未成年人和成年人归组</h5><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_436534" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div><div class="line number3 index2 alt2">3</div><div class="line number4 index3 alt1">4</div><div class="line number5 index4 alt2">5</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Map<</code><code class="htmlscript plain">Boolean</code><code class="htmlscript plain">, List<Person>> children = Stream.generate(new PersonSupplier()).</code></div><div class="line number2 index1 alt1"><code class="htmlscript spaces"> </code><code class="htmlscript plain">limit(100).</code></div><div class="line number3 index2 alt2"><code class="htmlscript spaces"> </code><code class="htmlscript plain">collect(Collectors.partitioningBy(p -> p.getAge() < 18));</code></div><div class="line number4 index3 alt1"><code class="htmlscript plain">System.out.println("Children number: " + children.get(true).size());</code></div><div class="line number5 index4 alt2"><code class="htmlscript plain">System.out.println("Adult number: " + children.get(false).size());</code></div></div></td></tr></tbody></table></div></div></div></span><p>输出结果：</p><span class="dw-code-nohighlight"><div class="ibm-syntax-container"><div><div id="highlighter_795042" class="syntaxhighlighter htmlscript"><table border="0" cellpadding="0" cellspacing="0" role="none"><tbody><tr><td class="gutter"><div class="line number1 index0 alt2">1</div><div class="line number2 index1 alt1">2</div></td><td class="code"><div class="container"><div class="line number1 index0 alt2"><code class="htmlscript plain">Children number: 23 </code></div><div class="line number2 index1 alt1"><code class="htmlscript plain">Adult number: 77</code></div></div></td></tr></tbody></table></div></div></div></span><p>在使用条件“年龄小于 18”进行分组后可以看到，不到 18 岁的未成年人是一组，成年人是另外一组。partitioningBy 其实是一种特殊的
groupingBy，它依照条件测试的是否两种结果来构造返回的数据结构，get(true) 和 get(false) 能即为全部的元素对象。</p><h2 id="major11" class="ibm-h2">结束语</h2><p>总之，Stream 的特性可以归纳为：</p><ul class="ibm-bullet-list"><li> 不是数据结构</li></ul><ul class="ibm-bullet-list"><li> 它没有内部存储，它只是用操作管道从 source（数据结构、数组、generator function、IO channel）抓取数据。</li><li> 它也绝不修改自己所封装的底层数据结构的数据。例如 Stream 的 filter 操作会产生一个不包含被过滤元素的新 Stream，而不是从 source 删除那些元素。</li></ul><ul class="ibm-bullet-list"><li> 所有 Stream 的操作必须以 lambda 表达式为参数</li><li> 不支持索引访问</li></ul><ul class="ibm-bullet-list"><li> 你可以请求第一个元素，但无法请求第二个，第三个，或最后一个。不过请参阅下一项。</li></ul><ul class="ibm-bullet-list"><li> 很容易生成数组或者 List</li><li> 惰性化</li></ul><ul class="ibm-bullet-list"><li> 很多 Stream 操作是向后延迟的，一直到它弄清楚了最后需要多少数据才会开始。</li><li> Intermediate 操作永远是惰性化的。</li></ul><ul class="ibm-bullet-list"><li> 并行能力</li></ul><ul class="ibm-bullet-list"><li> 当一个 Stream 是并行化的，就不需要再写多线程代码，所有对它的操作会自动并行进行的。</li></ul><ul class="ibm-bullet-list"><li> 可以是无限的<ul class="ibm-bullet-list"><li> 集合有固定大小，Stream 则不必。limit(n) 和 findFirst() 这类的 short-circuiting 操作可以对无限的 Stream
进行运算并很快完成。</li></ul></li></ul><!--CMA ID: 983042--><!--Site ID: 10--><!--XSLT stylesheet used to transform this file: dw-document-html-8.0.xsl-->