001/*
002 * Copyright (C) 2007 The Guava Authors
003 *
004 * Licensed under the Apache License, Version 2.0 (the "License");
005 * you may not use this file except in compliance with the License.
006 * You may obtain a copy of the License at
007 *
008 * http://www.apache.org/licenses/LICENSE-2.0
009 *
010 * Unless required by applicable law or agreed to in writing, software
011 * distributed under the License is distributed on an "AS IS" BASIS,
012 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
013 * See the License for the specific language governing permissions and
014 * limitations under the License.
015 */
016
017package com.google.common.collect;
018
019import com.google.common.annotations.GwtCompatible;
020
021import java.util.Collection;
022import java.util.Collections;
023import java.util.Iterator;
024import java.util.List;
025import java.util.Set;
026
027import javax.annotation.Nullable;
028
029/**
030 * A collection that supports order-independent equality, like {@link Set}, but
031 * may have duplicate elements. A multiset is also sometimes called a
032 * <i>bag</i>.
033 *
034 * <p>Elements of a multiset that are equal to one another are referred to as
035 * <i>occurrences</i> of the same single element. The total number of
036 * occurrences of an element in a multiset is called the <i>count</i> of that
037 * element (the terms "frequency" and "multiplicity" are equivalent, but not
038 * used in this API). Since the count of an element is represented as an {@code
039 * int}, a multiset may never contain more than {@link Integer#MAX_VALUE}
040 * occurrences of any one element.
041 *
042 * <p>{@code Multiset} refines the specifications of several methods from
043 * {@code Collection}. It also defines an additional query operation, {@link
044 * #count}, which returns the count of an element. There are five new
045 * bulk-modification operations, for example {@link #add(Object, int)}, to add
046 * or remove multiple occurrences of an element at once, or to set the count of
047 * an element to a specific value. These modification operations are optional,
048 * but implementations which support the standard collection operations {@link
049 * #add(Object)} or {@link #remove(Object)} are encouraged to implement the
050 * related methods as well. Finally, two collection views are provided: {@link
051 * #elementSet} contains the distinct elements of the multiset "with duplicates
052 * collapsed", and {@link #entrySet} is similar but contains {@link Entry
053 * Multiset.Entry} instances, each providing both a distinct element and the
054 * count of that element.
055 *
056 * <p>In addition to these required methods, implementations of {@code
057 * Multiset} are expected to provide two {@code static} creation methods:
058 * {@code create()}, returning an empty multiset, and {@code
059 * create(Iterable<? extends E>)}, returning a multiset containing the
060 * given initial elements. This is simply a refinement of {@code Collection}'s
061 * constructor recommendations, reflecting the new developments of Java 5.
062 *
063 * <p>As with other collection types, the modification operations are optional,
064 * and should throw {@link UnsupportedOperationException} when they are not
065 * implemented. Most implementations should support either all add operations
066 * or none of them, all removal operations or none of them, and if and only if
067 * all of these are supported, the {@code setCount} methods as well.
068 *
069 * <p>A multiset uses {@link Object#equals} to determine whether two instances
070 * should be considered "the same," <i>unless specified otherwise</i> by the
071 * implementation.
072 *
073 * <p>Common implementations include {@link ImmutableMultiset}, {@link
074 * HashMultiset}, and {@link ConcurrentHashMultiset}.
075 *
076 * <p>If your values may be zero, negative, or outside the range of an int, you
077 * may wish to use {@link com.google.common.util.concurrent.AtomicLongMap}
078 * instead. Note, however, that unlike {@code Multiset}, {@code AtomicLongMap}
079 * does not automatically remove zeros.
080 *
081 * <p>See the Guava User Guide article on <a href=
082 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#multiset">
083 * {@code Multiset}</a>.
084 *
085 * @author Kevin Bourrillion
086 * @since 2.0
087 */
088@GwtCompatible
089public interface Multiset<E> extends Collection<E> {
090  // Query Operations
091
092  /**
093   * Returns the number of occurrences of an element in this multiset (the
094   * <i>count</i> of the element). Note that for an {@link Object#equals}-based
095   * multiset, this gives the same result as {@link Collections#frequency}
096   * (which would presumably perform more poorly).
097   *
098   * <p><b>Note:</b> the utility method {@link Iterables#frequency} generalizes
099   * this operation; it correctly delegates to this method when dealing with a
100   * multiset, but it can also accept any other iterable type.
101   *
102   * @param element the element to count occurrences of
103   * @return the number of occurrences of the element in this multiset; possibly
104   *     zero but never negative
105   */
106  int count(@Nullable Object element);
107
108  // Bulk Operations
109
110  /**
111   * Adds a number of occurrences of an element to this multiset. Note that if
112   * {@code occurrences == 1}, this method has the identical effect to {@link
113   * #add(Object)}. This method is functionally equivalent (except in the case
114   * of overflow) to the call {@code addAll(Collections.nCopies(element,
115   * occurrences))}, which would presumably perform much more poorly.
116   *
117   * @param element the element to add occurrences of; may be null only if
118   *     explicitly allowed by the implementation
119   * @param occurrences the number of occurrences of the element to add. May be
120   *     zero, in which case no change will be made.
121   * @return the count of the element before the operation; possibly zero
122   * @throws IllegalArgumentException if {@code occurrences} is negative, or if
123   *     this operation would result in more than {@link Integer#MAX_VALUE}
124   *     occurrences of the element
125   * @throws NullPointerException if {@code element} is null and this
126   *     implementation does not permit null elements. Note that if {@code
127   *     occurrences} is zero, the implementation may opt to return normally.
128   */
129  int add(@Nullable E element, int occurrences);
130
131  /**
132   * Removes a number of occurrences of the specified element from this
133   * multiset. If the multiset contains fewer than this number of occurrences to
134   * begin with, all occurrences will be removed.  Note that if
135   * {@code occurrences == 1}, this is functionally equivalent to the call
136   * {@code remove(element)}.
137   *
138   * @param element the element to conditionally remove occurrences of
139   * @param occurrences the number of occurrences of the element to remove. May
140   *     be zero, in which case no change will be made.
141   * @return the count of the element before the operation; possibly zero
142   * @throws IllegalArgumentException if {@code occurrences} is negative
143   */
144  int remove(@Nullable Object element, int occurrences);
145
146  /**
147   * Adds or removes the necessary occurrences of an element such that the
148   * element attains the desired count.
149   *
150   * @param element the element to add or remove occurrences of; may be null
151   *     only if explicitly allowed by the implementation
152   * @param count the desired count of the element in this multiset
153   * @return the count of the element before the operation; possibly zero
154   * @throws IllegalArgumentException if {@code count} is negative
155   * @throws NullPointerException if {@code element} is null and this
156   *     implementation does not permit null elements. Note that if {@code
157   *     count} is zero, the implementor may optionally return zero instead.
158   */
159  int setCount(E element, int count);
160
161  /**
162   * Conditionally sets the count of an element to a new value, as described in
163   * {@link #setCount(Object, int)}, provided that the element has the expected
164   * current count. If the current count is not {@code oldCount}, no change is
165   * made.
166   *
167   * @param element the element to conditionally set the count of; may be null
168   *     only if explicitly allowed by the implementation
169   * @param oldCount the expected present count of the element in this multiset
170   * @param newCount the desired count of the element in this multiset
171   * @return {@code true} if the condition for modification was met. This
172   *     implies that the multiset was indeed modified, unless
173   *     {@code oldCount == newCount}.
174   * @throws IllegalArgumentException if {@code oldCount} or {@code newCount} is
175   *     negative
176   * @throws NullPointerException if {@code element} is null and the
177   *     implementation does not permit null elements. Note that if {@code
178   *     oldCount} and {@code newCount} are both zero, the implementor may
179   *     optionally return {@code true} instead.
180   */
181  boolean setCount(E element, int oldCount, int newCount);
182
183  // Views
184
185  /**
186   * Returns the set of distinct elements contained in this multiset. The
187   * element set is backed by the same data as the multiset, so any change to
188   * either is immediately reflected in the other. The order of the elements in
189   * the element set is unspecified.
190   *
191   * <p>If the element set supports any removal operations, these necessarily
192   * cause <b>all</b> occurrences of the removed element(s) to be removed from
193   * the multiset. Implementations are not expected to support the add
194   * operations, although this is possible.
195   *
196   * <p>A common use for the element set is to find the number of distinct
197   * elements in the multiset: {@code elementSet().size()}.
198   *
199   * @return a view of the set of distinct elements in this multiset
200   */
201  Set<E> elementSet();
202
203  /**
204   * Returns a view of the contents of this multiset, grouped into {@code
205   * Multiset.Entry} instances, each providing an element of the multiset and
206   * the count of that element. This set contains exactly one entry for each
207   * distinct element in the multiset (thus it always has the same size as the
208   * {@link #elementSet}). The order of the elements in the element set is
209   * unspecified.
210   *
211   * <p>The entry set is backed by the same data as the multiset, so any change
212   * to either is immediately reflected in the other. However, multiset changes
213   * may or may not be reflected in any {@code Entry} instances already
214   * retrieved from the entry set (this is implementation-dependent).
215   * Furthermore, implementations are not required to support modifications to
216   * the entry set at all, and the {@code Entry} instances themselves don't
217   * even have methods for modification. See the specific implementation class
218   * for more details on how its entry set handles modifications.
219   *
220   * @return a set of entries representing the data of this multiset
221   */
222  Set<Entry<E>> entrySet();
223
224  /**
225   * An unmodifiable element-count pair for a multiset. The {@link
226   * Multiset#entrySet} method returns a view of the multiset whose elements
227   * are of this class. A multiset implementation may return Entry instances
228   * that are either live "read-through" views to the Multiset, or immutable
229   * snapshots. Note that this type is unrelated to the similarly-named type
230   * {@code Map.Entry}.
231   *
232   * @since 2.0
233   */
234  interface Entry<E> {
235
236    /**
237     * Returns the multiset element corresponding to this entry. Multiple calls
238     * to this method always return the same instance.
239     *
240     * @return the element corresponding to this entry
241     */
242    E getElement();
243
244    /**
245     * Returns the count of the associated element in the underlying multiset.
246     * This count may either be an unchanging snapshot of the count at the time
247     * the entry was retrieved, or a live view of the current count of the
248     * element in the multiset, depending on the implementation. Note that in
249     * the former case, this method can never return zero, while in the latter,
250     * it will return zero if all occurrences of the element were since removed
251     * from the multiset.
252     *
253     * @return the count of the element; never negative
254     */
255    int getCount();
256
257    /**
258     * {@inheritDoc}
259     *
260     * <p>Returns {@code true} if the given object is also a multiset entry and
261     * the two entries represent the same element and count. That is, two
262     * entries {@code a} and {@code b} are equal if: <pre>   {@code
263     *
264     *   Objects.equal(a.getElement(), b.getElement())
265     *       && a.getCount() == b.getCount()}</pre>
266     */
267    @Override
268    // TODO(kevinb): check this wrt TreeMultiset?
269    boolean equals(Object o);
270
271    /**
272     * {@inheritDoc}
273     *
274     * <p>The hash code of a multiset entry for element {@code element} and
275     * count {@code count} is defined as: <pre>   {@code
276     *
277     *   ((element == null) ? 0 : element.hashCode()) ^ count}</pre>
278     */
279    @Override
280    int hashCode();
281
282    /**
283     * Returns the canonical string representation of this entry, defined as
284     * follows. If the count for this entry is one, this is simply the string
285     * representation of the corresponding element. Otherwise, it is the string
286     * representation of the element, followed by the three characters {@code
287     * " x "} (space, letter x, space), followed by the count.
288     */
289    @Override
290    String toString();
291  }
292
293  // Comparison and hashing
294
295  /**
296   * Compares the specified object with this multiset for equality. Returns
297   * {@code true} if the given object is also a multiset and contains equal
298   * elements with equal counts, regardless of order.
299   */
300  @Override
301  // TODO(kevinb): caveats about equivalence-relation?
302  boolean equals(@Nullable Object object);
303
304  /**
305   * Returns the hash code for this multiset. This is defined as the sum of
306   * <pre>   {@code
307   *
308   *   ((element == null) ? 0 : element.hashCode()) ^ count(element)}</pre>
309   *
310   * <p>over all distinct elements in the multiset. It follows that a multiset and
311   * its entry set always have the same hash code.
312   */
313  @Override
314  int hashCode();
315
316  /**
317   * {@inheritDoc}
318   *
319   * <p>It is recommended, though not mandatory, that this method return the
320   * result of invoking {@link #toString} on the {@link #entrySet}, yielding a
321   * result such as {@code [a x 3, c, d x 2, e]}.
322   */
323  @Override
324  String toString();
325
326  // Refined Collection Methods
327
328  /**
329   * {@inheritDoc}
330   *
331   * <p>Elements that occur multiple times in the multiset will appear
332   * multiple times in this iterator, though not necessarily sequentially.
333   */
334  @Override
335  Iterator<E> iterator();
336
337  /**
338   * Determines whether this multiset contains the specified element.
339   *
340   * <p>This method refines {@link Collection#contains} to further specify that
341   * it <b>may not</b> throw an exception in response to {@code element} being
342   * null or of the wrong type.
343   *
344   * @param element the element to check for
345   * @return {@code true} if this multiset contains at least one occurrence of
346   *     the element
347   */
348  @Override
349  boolean contains(@Nullable Object element);
350
351  /**
352   * Returns {@code true} if this multiset contains at least one occurrence of
353   * each element in the specified collection.
354   *
355   * <p>This method refines {@link Collection#containsAll} to further specify
356   * that it <b>may not</b> throw an exception in response to any of {@code
357   * elements} being null or of the wrong type.
358   *
359   * <p><b>Note:</b> this method does not take into account the occurrence
360   * count of an element in the two collections; it may still return {@code
361   * true} even if {@code elements} contains several occurrences of an element
362   * and this multiset contains only one. This is no different than any other
363   * collection type like {@link List}, but it may be unexpected to the user of
364   * a multiset.
365   *
366   * @param elements the collection of elements to be checked for containment in
367   *     this multiset
368   * @return {@code true} if this multiset contains at least one occurrence of
369   *     each element contained in {@code elements}
370   * @throws NullPointerException if {@code elements} is null
371   */
372  @Override
373  boolean containsAll(Collection<?> elements);
374
375  /**
376   * Adds a single occurrence of the specified element to this multiset.
377   *
378   * <p>This method refines {@link Collection#add}, which only <i>ensures</i>
379   * the presence of the element, to further specify that a successful call must
380   * always increment the count of the element, and the overall size of the
381   * collection, by one.
382   *
383   * <p>To both add the element and obtain the previous count of that element,
384   * use {@link #add(E, int) add}{@code (element, 1)} instead.
385   *
386   * @param element the element to add one occurrence of; may be null only if
387   *     explicitly allowed by the implementation
388   * @return {@code true} always, since this call is required to modify the
389   *     multiset, unlike other {@link Collection} types
390   * @throws NullPointerException if {@code element} is null and this
391   *     implementation does not permit null elements
392   * @throws IllegalArgumentException if {@link Integer#MAX_VALUE} occurrences
393   *     of {@code element} are already contained in this multiset
394   */
395  @Override
396  boolean add(E element);
397
398  /**
399   * Removes a <i>single</i> occurrence of the specified element from this
400   * multiset, if present.
401   *
402   * <p>This method refines {@link Collection#remove} to further specify that it
403   * <b>may not</b> throw an exception in response to {@code element} being null
404   * or of the wrong type.
405   *
406   * <p>To both remove the element and obtain the previous count of that element,
407   * use {@link #remove(E, int) remove}{@code (element, 1)} instead.
408   *
409   * @param element the element to remove one occurrence of
410   * @return {@code true} if an occurrence was found and removed
411   */
412  @Override
413  boolean remove(@Nullable Object element);
414
415  /**
416   * {@inheritDoc}
417   *
418   * <p><b>Note:</b> This method ignores how often any element might appear in
419   * {@code c}, and only cares whether or not an element appears at all.
420   * If you wish to remove one occurrence in this multiset for every occurrence
421   * in {@code c}, see {@link Multisets#removeOccurrences(Multiset, Multiset)}.
422   *
423   * <p>This method refines {@link Collection#removeAll} to further specify that
424   * it <b>may not</b> throw an exception in response to any of {@code elements}
425   * being null or of the wrong type.
426   */
427  @Override
428  boolean removeAll(Collection<?> c);
429
430  /**
431   * {@inheritDoc}
432   *
433   * <p><b>Note:</b> This method ignores how often any element might appear in
434   * {@code c}, and only cares whether or not an element appears at all.
435   * If you wish to remove one occurrence in this multiset for every occurrence
436   * in {@code c}, see {@link Multisets#retainOccurrences(Multiset, Multiset)}.
437   *
438   * <p>This method refines {@link Collection#retainAll} to further specify that
439   * it <b>may not</b> throw an exception in response to any of {@code elements}
440   * being null or of the wrong type.
441   *
442   * @see Multisets#retainOccurrences(Multiset, Multiset)
443   */
444  @Override
445  boolean retainAll(Collection<?> c);
446}