- A+
之前有写过ArrayList与LinkedList，分别用到了数组与链表的结构，下面来看看HashMap，了解其实现及扩容机制。
构造函数
HashMap有四个构造函数，其中比较有代表性的是下面这个
|
1 2 3 4 5 6 7 8 9 10 11 |
<span class="hljs-function"><span class="hljs-keyword">public</span> <span class="hljs-title">HashMap</span><span class="hljs-params">(<span class="hljs-keyword">int</span> initialCapacity, <span class="hljs-keyword">float</span> loadFactor)</span> </span>{ <span class="hljs-keyword">if</span> (initialCapacity < <span class="hljs-number">0</span>) <span class="hljs-keyword">throw</span> <span class="hljs-keyword">new</span> IllegalArgumentException(<span class="hljs-string">"Illegal initial capacity: "</span> + initialCapacity); <span class="hljs-keyword">if</span> (initialCapacity > MAXIMUM_CAPACITY) initialCapacity = MAXIMUM_CAPACITY; <span class="hljs-keyword">if</span> (loadFactor <= <span class="hljs-number">0</span> || Float.isNaN(loadFactor)) <span class="hljs-keyword">throw</span> <span class="hljs-keyword">new</span> IllegalArgumentException(<span class="hljs-string">"Illegal load factor: "</span> + loadFactor); <span class="hljs-keyword">this</span>.loadFactor = loadFactor; <span class="hljs-keyword">this</span>.threshold = tableSizeFor(initialCapacity); } |
需要注意的有两个入参initialCapacity和loadFactor，分别指HashMap的深度及负载因子（了解HashMap的实现后就能理解这两个值的含义）
在类中我们可以发现有两个常量，即他们的默认值，数组长度默认为16，当数组存储量达到16 * 0.75 = 12时，会进行扩容
|
1 2 3 |
<span class="hljs-keyword">static</span> <span class="hljs-keyword">final</span> <span class="hljs-keyword">int</span> DEFAULT_INITIAL_CAPACITY = <span class="hljs-number">1</span> << <span class="hljs-number">4</span>; <span class="hljs-keyword">static</span> <span class="hljs-keyword">final</span> <span class="hljs-keyword">float</span> DEFAULT_LOAD_FACTOR = <span class="hljs-number">0.75f</span>; |
再看看tableSizeFor(initialCapacity)
|
1 2 3 4 5 6 7 8 9 10 |
<span class="hljs-function"><span class="hljs-keyword">static</span> <span class="hljs-keyword">final</span> <span class="hljs-keyword">int</span> <span class="hljs-title">tableSizeFor</span><span class="hljs-params">(<span class="hljs-keyword">int</span> cap)</span> </span>{ <span class="hljs-keyword">int</span> n = cap - <span class="hljs-number">1</span>; n |= n >>> <span class="hljs-number">1</span>; n |= n >>> <span class="hljs-number">2</span>; n |= n >>> <span class="hljs-number">4</span>; n |= n >>> <span class="hljs-number">8</span>; n |= n >>> <span class="hljs-number">16</span>; <span class="hljs-keyword">return</span> (n < <span class="hljs-number">0</span>) ? <span class="hljs-number">1</span> : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + <span class="hljs-number">1</span>; } |
这里用于根据capacity生成最终的数组长度，并确保其为2的幂
数据结构
|
1 2 3 4 5 6 7 8 9 10 11 |
<span class="hljs-keyword">static</span> <span class="hljs-class"><span class="hljs-keyword">class</span> <span class="hljs-title">Node</span><K,V> <span class="hljs-title">implements</span> <span class="hljs-title">Map</span>.<span class="hljs-title">Entry</span><K,V> {</span> ... Node(<span class="hljs-keyword">int</span> hash, K key, V value, Node<K,V> next) { <span class="hljs-keyword">this</span>.hash = hash; <span class="hljs-keyword">this</span>.key = key; <span class="hljs-keyword">this</span>.value = value; <span class="hljs-keyword">this</span>.next = next; } ... } |
|
1 2 |
<span class="hljs-keyword">transient</span> Node<K,V>[] table; |
在HashMap中，可以发现上述两段代码，前一段代码是一个单向链表，可见，HashMap是一个链表数组（链表散列）
散列、put及get
再看看HashMap是如何实现存储的
|
1 2 3 4 |
<span class="hljs-function"><span class="hljs-keyword">public</span> V <span class="hljs-title">put</span><span class="hljs-params">(K key, V value)</span> </span>{ <span class="hljs-keyword">return</span> putVal(hash(key), key, value, <span class="hljs-literal">false</span>, <span class="hljs-literal">true</span>); } |
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 |
<span class="hljs-function"><span class="hljs-keyword">final</span> V <span class="hljs-title">putVal</span><span class="hljs-params">(<span class="hljs-keyword">int</span> hash, K key, V value, <span class="hljs-keyword">boolean</span> onlyIfAbsent, <span class="hljs-keyword">boolean</span> evict)</span> </span>{ Node<K,V>[] tab; Node<K,V> p; <span class="hljs-keyword">int</span> n, i; <span class="hljs-keyword">if</span> ((tab = table) == <span class="hljs-keyword">null</span> || (n = tab.length) == <span class="hljs-number">0</span>) n = (tab = resize()).length; <span class="hljs-keyword">if</span> ((p = tab[i = (n - <span class="hljs-number">1</span>) & hash]) == <span class="hljs-keyword">null</span>) tab[i] = newNode(hash, key, value, <span class="hljs-keyword">null</span>); <span class="hljs-keyword">else</span> { Node<K,V> e; K k; <span class="hljs-keyword">if</span> (p.hash == hash && ((k = p.key) == key || (key != <span class="hljs-keyword">null</span> && key.equals(k)))) e = p; <span class="hljs-keyword">else</span> <span class="hljs-keyword">if</span> (p <span class="hljs-keyword">instanceof</span> TreeNode) e = ((TreeNode<K,V>)p).putTreeVal(<span class="hljs-keyword">this</span>, tab, hash, key, value); <span class="hljs-keyword">else</span> { <span class="hljs-keyword">for</span> (<span class="hljs-keyword">int</span> binCount = <span class="hljs-number">0</span>; ; ++binCount) { <span class="hljs-keyword">if</span> ((e = p.next) == <span class="hljs-keyword">null</span>) { p.next = newNode(hash, key, value, <span class="hljs-keyword">null</span>); <span class="hljs-keyword">if</span> (binCount >= TREEIFY_THRESHOLD - <span class="hljs-number">1</span>) <span class="hljs-comment">// -1 for 1st</span> treeifyBin(tab, hash); <span class="hljs-keyword">break</span>; } <span class="hljs-keyword">if</span> (e.hash == hash && ((k = e.key) == key || (key != <span class="hljs-keyword">null</span> && key.equals(k)))) <span class="hljs-keyword">break</span>; p = e; } } <span class="hljs-keyword">if</span> (e != <span class="hljs-keyword">null</span>) { <span class="hljs-comment">// existing mapping for key</span> V oldValue = e.value; <span class="hljs-keyword">if</span> (!onlyIfAbsent || oldValue == <span class="hljs-keyword">null</span>) e.value = value; afterNodeAccess(e); <span class="hljs-keyword">return</span> oldValue; } } ++modCount; <span class="hljs-keyword">if</span> (++size > threshold) resize(); afterNodeInsertion(evict); <span class="hljs-keyword">return</span> <span class="hljs-keyword">null</span>; } |
源码比较长，逻辑上主要是
- 确定数组长度
- 根据hash值找数组位置 i = (n - 1) & hash
- 如果 i 位置为空，则新建Node并返回
- 如果 i 位置不为空，则判断key是否跟已存在的key一致（equals比较），一致则修改该Node为新建的Node，不一致则新建Node，并将Next指向新Node（HashMap认为最新加入的值为使用概率最大，故在链表前端增加Node，而不在尾端增加）
5.考虑是否需要扩容，根据前面说到的扩展因子进行判断，如何扩容后面再说
其中table位置使用了i = (n - 1) & hash，前面的tableSizeFor保证了n-1为全1的二进制数，通过&运算，取hash的后N位
hash的实现
|
1 2 3 4 5 |
<span class="hljs-function"><span class="hljs-keyword">static</span> <span class="hljs-keyword">final</span> <span class="hljs-keyword">int</span> <span class="hljs-title">hash</span><span class="hljs-params">(Object key)</span> </span>{ <span class="hljs-keyword">int</span> h; <span class="hljs-keyword">return</span> (key == <span class="hljs-keyword">null</span>) ? <span class="hljs-number">0</span> : (h = key.hashCode()) ^ (h >>> <span class="hljs-number">16</span>); } |
在hashCode后还有一个h>>>16异或的操作，hashCode的高16位与低16位异或
(n - 1) & hash
上图是 i = (n - 1) & hash的完整版，HashMap这样做，HashCode低位高位都利用起来，在低位间接的保留了高位的信息，加大低位的随机性，起到扰动函数的作用，减少hash冲突的几率，让HashMap的值尽量平均的分布在Node数组中
HashMap的取值与存储类似，关键代码依然是(n - 1) & hash，在遇到hash冲突时，则会进入链表取值，这个留到后面说
hash冲突
尽管HashMap的hash进行了高位低位混用的优化，但hash冲突是难免的，HashMap使用了链表的方式解决hash冲突
HashMap存储结构
从上图可以发现，当发生hash冲突时，HashMap使用了链表，之前贴出的put(key, value)方法中有一行是这样的
|
1 2 3 |
// 之前已经赋值 p = tab[i = (n - 1) & <span class="hljs-built_in">hash</span>]) <span class="hljs-keyword">if</span> (p.hash == <span class="hljs-built_in">hash</span> && ((k = p.key) == key || (key != null && key.equals(k)))) |
即在比较完hash值以后，会继续比较key是否一致，这样确保能找到正确的key
总结
HashMap采用了数组+链表的数据结构，利用hash()方法生成数组下标，有助于实现快速查找，而链表则解决了hash冲突的问题。而在扩容等方面，通过tableSizeFor，能完美的使用位运算，运算效率得到不错的提升。
ps: 近一个月都比较忙，工作+家庭，这篇文章断断续续写了三四次，从17年写到了18年，总算完了，不知道中间会不会有什么遗漏或者错误...
我的微信公众号
我的微信公众号扫一扫
