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