| LinkedHashMap.java |
1 /*
2 * %W% %E%
3 *
4 * Copyright (c) 2006, Oracle and/or its affiliates. All rights reserved.
5 * ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
6 */
7
8 package java.util;
9 import java.io.*;
10
11 /**
12 * <p>Hash table and linked list implementation of the <tt>Map</tt> interface,
13 * with predictable iteration order. This implementation differs from
14 * <tt>HashMap</tt> in that it maintains a doubly-linked list running through
15 * all of its entries. This linked list defines the iteration ordering,
16 * which is normally the order in which keys were inserted into the map
17 * (<i>insertion-order</i>). Note that insertion order is not affected
18 * if a key is <i>re-inserted</i> into the map. (A key <tt>k</tt> is
19 * reinserted into a map <tt>m</tt> if <tt>m.put(k, v)</tt> is invoked when
20 * <tt>m.containsKey(k)</tt> would return <tt>true</tt> immediately prior to
21 * the invocation.)
22 *
23 * <p>This implementation spares its clients from the unspecified, generally
24 * chaotic ordering provided by {@link HashMap} (and {@link Hashtable}),
25 * without incurring the increased cost associated with {@link TreeMap}. It
26 * can be used to produce a copy of a map that has the same order as the
27 * original, regardless of the original map's implementation:
28 * <pre>
29 * void foo(Map m) {
30 * Map copy = new LinkedHashMap(m);
31 * ...
32 * }
33 * </pre>
34 * This technique is particularly useful if a module takes a map on input,
35 * copies it, and later returns results whose order is determined by that of
36 * the copy. (Clients generally appreciate having things returned in the same
37 * order they were presented.)
38 *
39 * <p>A special {@link #LinkedHashMap(int,float,boolean) constructor} is
40 * provided to create a linked hash map whose order of iteration is the order
41 * in which its entries were last accessed, from least-recently accessed to
42 * most-recently (<i>access-order</i>). This kind of map is well-suited to
43 * building LRU caches. Invoking the <tt>put</tt> or <tt>get</tt> method
44 * results in an access to the corresponding entry (assuming it exists after
45 * the invocation completes). The <tt>putAll</tt> method generates one entry
46 * access for each mapping in the specified map, in the order that key-value
47 * mappings are provided by the specified map's entry set iterator. <i>No
48 * other methods generate entry accesses.</i> In particular, operations on
49 * collection-views do <i>not</i> affect the order of iteration of the backing
50 * map.
51 *
52 * <p>The {@link #removeEldestEntry(Map.Entry)} method may be overridden to
53 * impose a policy for removing stale mappings automatically when new mappings
54 * are added to the map.
55 *
56 * <p>This class provides all of the optional <tt>Map</tt> operations, and
57 * permits null elements. Like <tt>HashMap</tt>, it provides constant-time
58 * performance for the basic operations (<tt>add</tt>, <tt>contains</tt> and
59 * <tt>remove</tt>), assuming the hash function disperses elements
60 * properly among the buckets. Performance is likely to be just slightly
61 * below that of <tt>HashMap</tt>, due to the added expense of maintaining the
62 * linked list, with one exception: Iteration over the collection-views
63 * of a <tt>LinkedHashMap</tt> requires time proportional to the <i>size</i>
64 * of the map, regardless of its capacity. Iteration over a <tt>HashMap</tt>
65 * is likely to be more expensive, requiring time proportional to its
66 * <i>capacity</i>.
67 *
68 * <p>A linked hash map has two parameters that affect its performance:
69 * <i>initial capacity</i> and <i>load factor</i>. They are defined precisely
70 * as for <tt>HashMap</tt>. Note, however, that the penalty for choosing an
71 * excessively high value for initial capacity is less severe for this class
72 * than for <tt>HashMap</tt>, as iteration times for this class are unaffected
73 * by capacity.
74 *
75 * <p><strong>Note that this implementation is not synchronized.</strong>
76 * If multiple threads access a linked hash map concurrently, and at least
77 * one of the threads modifies the map structurally, it <em>must</em> be
78 * synchronized externally. This is typically accomplished by
79 * synchronizing on some object that naturally encapsulates the map.
80 *
81 * If no such object exists, the map should be "wrapped" using the
82 * {@link Collections#synchronizedMap Collections.synchronizedMap}
83 * method. This is best done at creation time, to prevent accidental
84 * unsynchronized access to the map:<pre>
85 * Map m = Collections.synchronizedMap(new LinkedHashMap(...));</pre>
86 *
87 * A structural modification is any operation that adds or deletes one or more
88 * mappings or, in the case of access-ordered linked hash maps, affects
89 * iteration order. In insertion-ordered linked hash maps, merely changing
90 * the value associated with a key that is already contained in the map is not
91 * a structural modification. <strong>In access-ordered linked hash maps,
92 * merely querying the map with <tt>get</tt> is a structural
93 * modification.</strong>)
94 *
95 * <p>The iterators returned by the <tt>iterator</tt> method of the collections
96 * returned by all of this class's collection view methods are
97 * <em>fail-fast</em>: if the map is structurally modified at any time after
98 * the iterator is created, in any way except through the iterator's own
99 * <tt>remove</tt> method, the iterator will throw a {@link
100 * ConcurrentModificationException}. Thus, in the face of concurrent
101 * modification, the iterator fails quickly and cleanly, rather than risking
102 * arbitrary, non-deterministic behavior at an undetermined time in the future.
103 *
104 * <p>Note that the fail-fast behavior of an iterator cannot be guaranteed
105 * as it is, generally speaking, impossible to make any hard guarantees in the
106 * presence of unsynchronized concurrent modification. Fail-fast iterators
107 * throw <tt>ConcurrentModificationException</tt> on a best-effort basis.
108 * Therefore, it would be wrong to write a program that depended on this
109 * exception for its correctness: <i>the fail-fast behavior of iterators
110 * should be used only to detect bugs.</i>
111 *
112 * <p>This class is a member of the
113 * <a href="{@docRoot}/../technotes/guides/collections/index.html">
114 * Java Collections Framework</a>.
115 *
116 * @param <K> the type of keys maintained by this map
117 * @param <V> the type of mapped values
118 *
119 * @author Josh Bloch
120 * @version %I%, %G%
121 * @see Object#hashCode()
122 * @see Collection
123 * @see Map
124 * @see HashMap
125 * @see TreeMap
126 * @see Hashtable
127 * @since 1.4
128 */
129
130 public class LinkedHashMap<K,V>
131 extends HashMap<K,V>
132 implements Map<K,V>
133 {
134
135 private static final long serialVersionUID = 3801124242820219131L;
136
137 /**
138 * The head of the doubly linked list.
139 */
140 private transient Entry<K,V> header;
141
142 /**
143 * The iteration ordering method for this linked hash map: <tt>true</tt>
144 * for access-order, <tt>false</tt> for insertion-order.
145 *
146 * @serial
147 */
148 private final boolean accessOrder;
149
150 /**
151 * Constructs an empty insertion-ordered <tt>LinkedHashMap</tt> instance
152 * with the specified initial capacity and load factor.
153 *
154 * @param initialCapacity the initial capacity
155 * @param loadFactor the load factor
156 * @throws IllegalArgumentException if the initial capacity is negative
157 * or the load factor is nonpositive
158 */
159 public LinkedHashMap(int initialCapacity, float loadFactor) {
160 super(initialCapacity, loadFactor);
161 accessOrder = false;
162 }
163
164 /**
165 * Constructs an empty insertion-ordered <tt>LinkedHashMap</tt> instance
166 * with the specified initial capacity and a default load factor (0.75).
167 *
168 * @param initialCapacity the initial capacity
169 * @throws IllegalArgumentException if the initial capacity is negative
170 */
171 public LinkedHashMap(int initialCapacity) {
172 super(initialCapacity);
173 accessOrder = false;
174 }
175
176 /**
177 * Constructs an empty insertion-ordered <tt>LinkedHashMap</tt> instance
178 * with the default initial capacity (16) and load factor (0.75).
179 */
180 public LinkedHashMap() {
181 super();
182 accessOrder = false;
183 }
184
185 /**
186 * Constructs an insertion-ordered <tt>LinkedHashMap</tt> instance with
187 * the same mappings as the specified map. The <tt>LinkedHashMap</tt>
188 * instance is created with a default load factor (0.75) and an initial
189 * capacity sufficient to hold the mappings in the specified map.
190 *
191 * @param m the map whose mappings are to be placed in this map
192 * @throws NullPointerException if the specified map is null
193 */
194 public LinkedHashMap(Map<? extends K, ? extends V> m) {
195 super(m);
196 accessOrder = false;
197 }
198
199 /**
200 * Constructs an empty <tt>LinkedHashMap</tt> instance with the
201 * specified initial capacity, load factor and ordering mode.
202 *
203 * @param initialCapacity the initial capacity
204 * @param loadFactor the load factor
205 * @param accessOrder the ordering mode - <tt>true</tt> for
206 * access-order, <tt>false</tt> for insertion-order
207 * @throws IllegalArgumentException if the initial capacity is negative
208 * or the load factor is nonpositive
209 */
210 public LinkedHashMap(int initialCapacity,
211 float loadFactor,
212 boolean accessOrder) {
213 super(initialCapacity, loadFactor);
214 this.accessOrder = accessOrder;
215 }
216
217 /**
218 * Called by superclass constructors and pseudoconstructors (clone,
219 * readObject) before any entries are inserted into the map. Initializes
220 * the chain.
221 */
222 void init() {
223 header = new Entry<K,V>(-1, null, null, null);
224 header.before = header.after = header;
225 }
226
227 /**
228 * Transfers all entries to new table array. This method is called
229 * by superclass resize. It is overridden for performance, as it is
230 * faster to iterate using our linked list.
231 */
232 void transfer(HashMap.Entry[] newTable) {
233 int newCapacity = newTable.length;
234 for (Entry<K,V> e = header.after; e != header; e = e.after) {
235 int index = indexFor(e.hash, newCapacity);
236 e.next = newTable[index];
237 newTable[index] = e;
238 }
239 }
240
241
242 /**
243 * Returns <tt>true</tt> if this map maps one or more keys to the
244 * specified value.
245 *
246 * @param value value whose presence in this map is to be tested
247 * @return <tt>true</tt> if this map maps one or more keys to the
248 * specified value
249 */
250 public boolean containsValue(Object value) {
251 // Overridden to take advantage of faster iterator
252 if (value==null) {
253 for (Entry e = header.after; e != header; e = e.after)
254 if (e.value==null)
255 return true;
256 } else {
257 for (Entry e = header.after; e != header; e = e.after)
258 if (value.equals(e.value))
259 return true;
260 }
261 return false;
262 }
263
264 /**
265 * Returns the value to which the specified key is mapped,
266 * or {@code null} if this map contains no mapping for the key.
267 *
268 * <p>More formally, if this map contains a mapping from a key
269 * {@code k} to a value {@code v} such that {@code (key==null ? k==null :
270 * key.equals(k))}, then this method returns {@code v}; otherwise
271 * it returns {@code null}. (There can be at most one such mapping.)
272 *
273 * <p>A return value of {@code null} does not <i>necessarily</i>
274 * indicate that the map contains no mapping for the key; it's also
275 * possible that the map explicitly maps the key to {@code null}.
276 * The {@link #containsKey containsKey} operation may be used to
277 * distinguish these two cases.
278 */
279 public V get(Object key) {
280 Entry<K,V> e = (Entry<K,V>)getEntry(key);
281 if (e == null)
282 return null;
283 e.recordAccess(this);
284 return e.value;
285 }
286
287 /**
288 * Removes all of the mappings from this map.
289 * The map will be empty after this call returns.
290 */
291 public void clear() {
292 super.clear();
293 header.before = header.after = header;
294 }
295
296 /**
297 * LinkedHashMap entry.
298 */
299 private static class Entry<K,V> extends HashMap.Entry<K,V> {
300 // These fields comprise the doubly linked list used for iteration.
301 Entry<K,V> before, after;
302
303 Entry(int hash, K key, V value, HashMap.Entry<K,V> next) {
304 super(hash, key, value, next);
305 }
306
307 /**
308 * Removes this entry from the linked list.
309 */
310 private void remove() {
311 before.after = after;
312 after.before = before;
313 }
314
315 /**
316 * Inserts this entry before the specified existing entry in the list.
317 */
318 private void addBefore(Entry<K,V> existingEntry) {
319 after = existingEntry;
320 before = existingEntry.before;
321 before.after = this;
322 after.before = this;
323 }
324
325 /**
326 * This method is invoked by the superclass whenever the value
327 * of a pre-existing entry is read by Map.get or modified by Map.set.
328 * If the enclosing Map is access-ordered, it moves the entry
329 * to the end of the list; otherwise, it does nothing.
330 */
331 void recordAccess(HashMap<K,V> m) {
332 LinkedHashMap<K,V> lm = (LinkedHashMap<K,V>)m;
333 if (lm.accessOrder) {
334 lm.modCount++;
335 remove();
336 addBefore(lm.header);
337 }
338 }
339
340 void recordRemoval(HashMap<K,V> m) {
341 remove();
342 }
343 }
344
345 private abstract class LinkedHashIterator<T> implements Iterator<T> {
346 Entry<K,V> nextEntry = header.after;
347 Entry<K,V> lastReturned = null;
348
349 /**
350 * The modCount value that the iterator believes that the backing
351 * List should have. If this expectation is violated, the iterator
352 * has detected concurrent modification.
353 */
354 int expectedModCount = modCount;
355
356 public boolean hasNext() {
357 return nextEntry != header;
358 }
359
360 public void remove() {
361 if (lastReturned == null)
362 throw new IllegalStateException();
363 if (modCount != expectedModCount)
364 throw new ConcurrentModificationException();
365
366 LinkedHashMap.this.remove(lastReturned.key);
367 lastReturned = null;
368 expectedModCount = modCount;
369 }
370
371 Entry<K,V> nextEntry() {
372 if (modCount != expectedModCount)
373 throw new ConcurrentModificationException();
374 if (nextEntry == header)
375 throw new NoSuchElementException();
376
377 Entry<K,V> e = lastReturned = nextEntry;
378 nextEntry = e.after;
379 return e;
380 }
381 }
382
383 private class KeyIterator extends LinkedHashIterator<K> {
384 public K next() { return nextEntry().getKey(); }
385 }
386
387 private class ValueIterator extends LinkedHashIterator<V> {
388 public V next() { return nextEntry().value; }
389 }
390
391 private class EntryIterator extends LinkedHashIterator<Map.Entry<K,V>> {
392 public Map.Entry<K,V> next() { return nextEntry(); }
393 }
394
395 // These Overrides alter the behavior of superclass view iterator() methods
396 Iterator<K> newKeyIterator() { return new KeyIterator(); }
397 Iterator<V> newValueIterator() { return new ValueIterator(); }
398 Iterator<Map.Entry<K,V>> newEntryIterator() { return new EntryIterator(); }
399
400 /**
401 * This override alters behavior of superclass put method. It causes newly
402 * allocated entry to get inserted at the end of the linked list and
403 * removes the eldest entry if appropriate.
404 */
405 void addEntry(int hash, K key, V value, int bucketIndex) {
406 createEntry(hash, key, value, bucketIndex);
407
408 // Remove eldest entry if instructed, else grow capacity if appropriate
409 Entry<K,V> eldest = header.after;
410 if (removeEldestEntry(eldest)) {
411 removeEntryForKey(eldest.key);
412 } else {
413 if (size >= threshold)
414 resize(2 * table.length);
415 }
416 }
417
418 /**
419 * This override differs from addEntry in that it doesn't resize the
420 * table or remove the eldest entry.
421 */
422 void createEntry(int hash, K key, V value, int bucketIndex) {
423 HashMap.Entry<K,V> old = table[bucketIndex];
424 Entry<K,V> e = new Entry<K,V>(hash, key, value, old);
425 table[bucketIndex] = e;
426 e.addBefore(header);
427 size++;
428 }
429
430 /**
431 * Returns <tt>true</tt> if this map should remove its eldest entry.
432 * This method is invoked by <tt>put</tt> and <tt>putAll</tt> after
433 * inserting a new entry into the map. It provides the implementor
434 * with the opportunity to remove the eldest entry each time a new one
435 * is added. This is useful if the map represents a cache: it allows
436 * the map to reduce memory consumption by deleting stale entries.
437 *
438 * <p>Sample use: this override will allow the map to grow up to 100
439 * entries and then delete the eldest entry each time a new entry is
440 * added, maintaining a steady state of 100 entries.
441 * <pre>
442 * private static final int MAX_ENTRIES = 100;
443 *
444 * protected boolean removeEldestEntry(Map.Entry eldest) {
445 * return size() > MAX_ENTRIES;
446 * }
447 * </pre>
448 *
449 * <p>This method typically does not modify the map in any way,
450 * instead allowing the map to modify itself as directed by its
451 * return value. It <i>is</i> permitted for this method to modify
452 * the map directly, but if it does so, it <i>must</i> return
453 * <tt>false</tt> (indicating that the map should not attempt any
454 * further modification). The effects of returning <tt>true</tt>
455 * after modifying the map from within this method are unspecified.
456 *
457 * <p>This implementation merely returns <tt>false</tt> (so that this
458 * map acts like a normal map - the eldest element is never removed).
459 *
460 * @param eldest The least recently inserted entry in the map, or if
461 * this is an access-ordered map, the least recently accessed
462 * entry. This is the entry that will be removed it this
463 * method returns <tt>true</tt>. If the map was empty prior
464 * to the <tt>put</tt> or <tt>putAll</tt> invocation resulting
465 * in this invocation, this will be the entry that was just
466 * inserted; in other words, if the map contains a single
467 * entry, the eldest entry is also the newest.
468 * @return <tt>true</tt> if the eldest entry should be removed
469 * from the map; <tt>false</tt> if it should be retained.
470 */
471 protected boolean removeEldestEntry(Map.Entry<K,V> eldest) {
472 return false;
473 }
474 }
475