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 static com.google.common.base.Preconditions.checkNotNull; 020 021import com.google.common.annotations.Beta; 022import com.google.common.annotations.GwtCompatible; 023import com.google.errorprone.annotations.CanIgnoreReturnValue; 024import com.google.errorprone.annotations.CompatibleWith; 025import java.util.Collection; 026import java.util.Collections; 027import java.util.Iterator; 028import java.util.List; 029import java.util.Set; 030import java.util.Spliterator; 031import java.util.function.Consumer; 032import java.util.function.ObjIntConsumer; 033import javax.annotation.CheckForNull; 034import org.checkerframework.checker.nullness.qual.Nullable; 035 036/** 037 * A collection that supports order-independent equality, like {@link Set}, but may have duplicate 038 * elements. A multiset is also sometimes called a <i>bag</i>. 039 * 040 * <p>Elements of a multiset that are equal to one another are referred to as <i>occurrences</i> of 041 * the same single element. The total number of occurrences of an element in a multiset is called 042 * the <i>count</i> of that element (the terms "frequency" and "multiplicity" are equivalent, but 043 * not used in this API). Since the count of an element is represented as an {@code int}, a multiset 044 * may never contain more than {@link Integer#MAX_VALUE} occurrences of any one element. 045 * 046 * <p>{@code Multiset} refines the specifications of several methods from {@code Collection}. It 047 * also defines an additional query operation, {@link #count}, which returns the count of an 048 * element. There are five new bulk-modification operations, for example {@link #add(Object, int)}, 049 * to add or remove multiple occurrences of an element at once, or to set the count of an element to 050 * a specific value. These modification operations are optional, but implementations which support 051 * the standard collection operations {@link #add(Object)} or {@link #remove(Object)} are encouraged 052 * to implement the related methods as well. Finally, two collection views are provided: {@link 053 * #elementSet} contains the distinct elements of the multiset "with duplicates collapsed", and 054 * {@link #entrySet} is similar but contains {@link Entry Multiset.Entry} instances, each providing 055 * both a distinct element and the count of that element. 056 * 057 * <p>In addition to these required methods, implementations of {@code Multiset} are expected to 058 * provide two {@code static} creation methods: {@code create()}, returning an empty multiset, and 059 * {@code create(Iterable<? extends E>)}, returning a multiset containing the given initial 060 * elements. This is simply a refinement of {@code Collection}'s constructor recommendations, 061 * reflecting the new developments of Java 5. 062 * 063 * <p>As with other collection types, the modification operations are optional, and should throw 064 * {@link UnsupportedOperationException} when they are not implemented. Most implementations should 065 * support either all add operations or none of them, all removal operations or none of them, and if 066 * and only if all of these are supported, the {@code setCount} methods as well. 067 * 068 * <p>A multiset uses {@link Object#equals} to determine whether two instances should be considered 069 * "the same," <i>unless specified otherwise</i> by the implementation. 070 * 071 * <p><b>Warning:</b> as with normal {@link Set}s, it is almost always a bad idea to modify an 072 * element (in a way that affects its {@link Object#equals} behavior) while it is contained in a 073 * multiset. Undefined behavior and bugs will result. 074 * 075 * <p>Common implementations include {@link ImmutableMultiset}, {@link HashMultiset}, and {@link 076 * ConcurrentHashMultiset}. 077 * 078 * <p>If your values may be zero, negative, or outside the range of an int, you may wish to use 079 * {@link com.google.common.util.concurrent.AtomicLongMap} instead. Note, however, that unlike 080 * {@code Multiset}, {@code AtomicLongMap} does not automatically remove zeros. 081 * 082 * <p>See the Guava User Guide article on <a href= 083 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#multiset">{@code Multiset}</a>. 084 * 085 * @author Kevin Bourrillion 086 * @since 2.0 087 */ 088@GwtCompatible 089@ElementTypesAreNonnullByDefault 090public interface Multiset<E extends @Nullable Object> extends Collection<E> { 091 // Query Operations 092 093 /** 094 * Returns the total number of all occurrences of all elements in this multiset. 095 * 096 * <p><b>Note:</b> this method does not return the number of <i>distinct elements</i> in the 097 * multiset, which is given by {@code entrySet().size()}. 098 */ 099 @Override 100 int size(); 101 102 /** 103 * Returns the number of occurrences of an element in this multiset (the <i>count</i> of the 104 * element). Note that for an {@link Object#equals}-based multiset, this gives the same result as 105 * {@link Collections#frequency} (which would presumably perform more poorly). 106 * 107 * <p><b>Note:</b> the utility method {@link Iterables#frequency} generalizes this operation; it 108 * correctly delegates to this method when dealing with a multiset, but it can also accept any 109 * other iterable type. 110 * 111 * @param element the element to count occurrences of 112 * @return the number of occurrences of the element in this multiset; possibly zero but never 113 * negative 114 */ 115 int count(@CompatibleWith("E") @CheckForNull Object element); 116 117 // Bulk Operations 118 119 /** 120 * Adds a number of occurrences of an element to this multiset. Note that if {@code occurrences == 121 * 1}, this method has the identical effect to {@link #add(Object)}. This method is functionally 122 * equivalent (except in the case of overflow) to the call {@code 123 * addAll(Collections.nCopies(element, occurrences))}, which would presumably perform much more 124 * poorly. 125 * 126 * @param element the element to add occurrences of; may be null only if explicitly allowed by the 127 * implementation 128 * @param occurrences the number of occurrences of the element to add. May be zero, in which case 129 * no change will be made. 130 * @return the count of the element before the operation; possibly zero 131 * @throws IllegalArgumentException if {@code occurrences} is negative, or if this operation would 132 * result in more than {@link Integer#MAX_VALUE} occurrences of the element 133 * @throws NullPointerException if {@code element} is null and this implementation does not permit 134 * null elements. Note that if {@code occurrences} is zero, the implementation may opt to 135 * return normally. 136 */ 137 @CanIgnoreReturnValue 138 int add(@ParametricNullness E element, int occurrences); 139 140 /** 141 * Adds a single occurrence of the specified element to this multiset. 142 * 143 * <p>This method refines {@link Collection#add}, which only <i>ensures</i> the presence of the 144 * element, to further specify that a successful call must always increment the count of the 145 * element, and the overall size of the collection, by one. 146 * 147 * <p>To both add the element and obtain the previous count of that element, use {@link 148 * #add(Object, int) add}{@code (element, 1)} instead. 149 * 150 * @param element the element to add one occurrence of; may be null only if explicitly allowed by 151 * the implementation 152 * @return {@code true} always, since this call is required to modify the multiset, unlike other 153 * {@link Collection} types 154 * @throws NullPointerException if {@code element} is null and this implementation does not permit 155 * null elements 156 * @throws IllegalArgumentException if {@link Integer#MAX_VALUE} occurrences of {@code element} 157 * are already contained in this multiset 158 */ 159 @CanIgnoreReturnValue 160 @Override 161 boolean add(@ParametricNullness E element); 162 163 /** 164 * Removes a number of occurrences of the specified element from this multiset. If the multiset 165 * contains fewer than this number of occurrences to begin with, all occurrences will be removed. 166 * Note that if {@code occurrences == 1}, this is functionally equivalent to the call {@code 167 * remove(element)}. 168 * 169 * @param element the element to conditionally remove occurrences of 170 * @param occurrences the number of occurrences of the element to remove. May be zero, in which 171 * case no change will be made. 172 * @return the count of the element before the operation; possibly zero 173 * @throws IllegalArgumentException if {@code occurrences} is negative 174 */ 175 @CanIgnoreReturnValue 176 int remove(@CompatibleWith("E") @CheckForNull Object element, int occurrences); 177 178 /** 179 * Removes a <i>single</i> occurrence of the specified element from this multiset, if present. 180 * 181 * <p>This method refines {@link Collection#remove} to further specify that it <b>may not</b> 182 * throw an exception in response to {@code element} being null or of the wrong type. 183 * 184 * <p>To both remove the element and obtain the previous count of that element, use {@link 185 * #remove(Object, int) remove}{@code (element, 1)} instead. 186 * 187 * @param element the element to remove one occurrence of 188 * @return {@code true} if an occurrence was found and removed 189 */ 190 @CanIgnoreReturnValue 191 @Override 192 boolean remove(@CheckForNull Object element); 193 194 /** 195 * Adds or removes the necessary occurrences of an element such that the element attains the 196 * desired count. 197 * 198 * @param element the element to add or remove occurrences of; may be null only if explicitly 199 * allowed by the implementation 200 * @param count the desired count of the element in this multiset 201 * @return the count of the element before the operation; possibly zero 202 * @throws IllegalArgumentException if {@code count} is negative 203 * @throws NullPointerException if {@code element} is null and this implementation does not permit 204 * null elements. Note that if {@code count} is zero, the implementor may optionally return 205 * zero instead. 206 */ 207 @CanIgnoreReturnValue 208 int setCount(@ParametricNullness E element, int count); 209 210 /** 211 * Conditionally sets the count of an element to a new value, as described in {@link 212 * #setCount(Object, int)}, provided that the element has the expected current count. If the 213 * current count is not {@code oldCount}, no change is made. 214 * 215 * @param element the element to conditionally set the count of; may be null only if explicitly 216 * allowed by the implementation 217 * @param oldCount the expected present count of the element in this multiset 218 * @param newCount the desired count of the element in this multiset 219 * @return {@code true} if the condition for modification was met. This implies that the multiset 220 * was indeed modified, unless {@code oldCount == newCount}. 221 * @throws IllegalArgumentException if {@code oldCount} or {@code newCount} is negative 222 * @throws NullPointerException if {@code element} is null and the implementation does not permit 223 * null elements. Note that if {@code oldCount} and {@code newCount} are both zero, the 224 * implementor may optionally return {@code true} instead. 225 */ 226 @CanIgnoreReturnValue 227 boolean setCount(@ParametricNullness E element, int oldCount, int newCount); 228 229 // Views 230 231 /** 232 * Returns the set of distinct elements contained in this multiset. The element set is backed by 233 * the same data as the multiset, so any change to either is immediately reflected in the other. 234 * The order of the elements in the element set is unspecified. 235 * 236 * <p>If the element set supports any removal operations, these necessarily cause <b>all</b> 237 * occurrences of the removed element(s) to be removed from the multiset. Implementations are not 238 * expected to support the add operations, although this is possible. 239 * 240 * <p>A common use for the element set is to find the number of distinct elements in the multiset: 241 * {@code elementSet().size()}. 242 * 243 * @return a view of the set of distinct elements in this multiset 244 */ 245 Set<E> elementSet(); 246 247 /** 248 * Returns a view of the contents of this multiset, grouped into {@code Multiset.Entry} instances, 249 * each providing an element of the multiset and the count of that element. This set contains 250 * exactly one entry for each distinct element in the multiset (thus it always has the same size 251 * as the {@link #elementSet}). The order of the elements in the element set is unspecified. 252 * 253 * <p>The entry set is backed by the same data as the multiset, so any change to either is 254 * immediately reflected in the other. However, multiset changes may or may not be reflected in 255 * any {@code Entry} instances already retrieved from the entry set (this is 256 * implementation-dependent). Furthermore, implementations are not required to support 257 * modifications to the entry set at all, and the {@code Entry} instances themselves don't even 258 * have methods for modification. See the specific implementation class for more details on how 259 * its entry set handles modifications. 260 * 261 * @return a set of entries representing the data of this multiset 262 */ 263 Set<Entry<E>> entrySet(); 264 265 /** 266 * An unmodifiable element-count pair for a multiset. The {@link Multiset#entrySet} method returns 267 * a view of the multiset whose elements are of this class. A multiset implementation may return 268 * Entry instances that are either live "read-through" views to the Multiset, or immutable 269 * snapshots. Note that this type is unrelated to the similarly-named type {@code Map.Entry}. 270 * 271 * @since 2.0 272 */ 273 interface Entry<E extends @Nullable Object> { 274 275 /** 276 * Returns the multiset element corresponding to this entry. Multiple calls to this method 277 * always return the same instance. 278 * 279 * @return the element corresponding to this entry 280 */ 281 @ParametricNullness 282 E getElement(); 283 284 /** 285 * Returns the count of the associated element in the underlying multiset. This count may either 286 * be an unchanging snapshot of the count at the time the entry was retrieved, or a live view of 287 * the current count of the element in the multiset, depending on the implementation. Note that 288 * in the former case, this method can never return zero, while in the latter, it will return 289 * zero if all occurrences of the element were since removed from the multiset. 290 * 291 * @return the count of the element; never negative 292 */ 293 int getCount(); 294 295 /** 296 * {@inheritDoc} 297 * 298 * <p>Returns {@code true} if the given object is also a multiset entry and the two entries 299 * represent the same element and count. That is, two entries {@code a} and {@code b} are equal 300 * if: 301 * 302 * <pre>{@code 303 * Objects.equal(a.getElement(), b.getElement()) 304 * && a.getCount() == b.getCount() 305 * }</pre> 306 */ 307 @Override 308 // TODO(kevinb): check this wrt TreeMultiset? 309 boolean equals(@CheckForNull Object o); 310 311 /** 312 * {@inheritDoc} 313 * 314 * <p>The hash code of a multiset entry for element {@code element} and count {@code count} is 315 * defined as: 316 * 317 * <pre>{@code 318 * ((element == null) ? 0 : element.hashCode()) ^ count 319 * }</pre> 320 */ 321 @Override 322 int hashCode(); 323 324 /** 325 * Returns the canonical string representation of this entry, defined as follows. If the count 326 * for this entry is one, this is simply the string representation of the corresponding element. 327 * Otherwise, it is the string representation of the element, followed by the three characters 328 * {@code " x "} (space, letter x, space), followed by the count. 329 */ 330 @Override 331 String toString(); 332 } 333 334 /** 335 * Runs the specified action for each distinct element in this multiset, and the number of 336 * occurrences of that element. For some {@code Multiset} implementations, this may be more 337 * efficient than iterating over the {@link #entrySet()} either explicitly or with {@code 338 * entrySet().forEach(action)}. 339 * 340 * @since 21.0 341 */ 342 @Beta 343 default void forEachEntry(ObjIntConsumer<? super E> action) { 344 checkNotNull(action); 345 entrySet().forEach(entry -> action.accept(entry.getElement(), entry.getCount())); 346 } 347 348 // Comparison and hashing 349 350 /** 351 * Compares the specified object with this multiset for equality. Returns {@code true} if the 352 * given object is also a multiset and contains equal elements with equal counts, regardless of 353 * order. 354 */ 355 @Override 356 // TODO(kevinb): caveats about equivalence-relation? 357 boolean equals(@CheckForNull Object object); 358 359 /** 360 * Returns the hash code for this multiset. This is defined as the sum of 361 * 362 * <pre>{@code 363 * ((element == null) ? 0 : element.hashCode()) ^ count(element) 364 * }</pre> 365 * 366 * <p>over all distinct elements in the multiset. It follows that a multiset and its entry set 367 * always have the same hash code. 368 */ 369 @Override 370 int hashCode(); 371 372 /** 373 * {@inheritDoc} 374 * 375 * <p>It is recommended, though not mandatory, that this method return the result of invoking 376 * {@link #toString} on the {@link #entrySet}, yielding a result such as {@code [a x 3, c, d x 2, 377 * e]}. 378 */ 379 @Override 380 String toString(); 381 382 // Refined Collection Methods 383 384 /** 385 * {@inheritDoc} 386 * 387 * <p>Elements that occur multiple times in the multiset will appear multiple times in this 388 * iterator, though not necessarily sequentially. 389 */ 390 @Override 391 Iterator<E> iterator(); 392 393 /** 394 * Determines whether this multiset contains the specified element. 395 * 396 * <p>This method refines {@link Collection#contains} to further specify that it <b>may not</b> 397 * throw an exception in response to {@code element} being null or of the wrong type. 398 * 399 * @param element the element to check for 400 * @return {@code true} if this multiset contains at least one occurrence of the element 401 */ 402 @Override 403 boolean contains(@CheckForNull Object element); 404 405 /** 406 * Returns {@code true} if this multiset contains at least one occurrence of each element in the 407 * specified collection. 408 * 409 * <p>This method refines {@link Collection#containsAll} to further specify that it <b>may not</b> 410 * throw an exception in response to any of {@code elements} being null or of the wrong type. 411 * 412 * <p><b>Note:</b> this method does not take into account the occurrence count of an element in 413 * the two collections; it may still return {@code true} even if {@code elements} contains several 414 * occurrences of an element and this multiset contains only one. This is no different than any 415 * other collection type like {@link List}, but it may be unexpected to the user of a multiset. 416 * 417 * @param elements the collection of elements to be checked for containment in this multiset 418 * @return {@code true} if this multiset contains at least one occurrence of each element 419 * contained in {@code elements} 420 * @throws NullPointerException if {@code elements} is null 421 */ 422 @Override 423 boolean containsAll(Collection<?> elements); 424 425 /** 426 * {@inheritDoc} 427 * 428 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 429 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 430 * multiset for every occurrence in {@code c}, see {@link Multisets#removeOccurrences(Multiset, 431 * Multiset)}. 432 * 433 * <p>This method refines {@link Collection#removeAll} to further specify that it <b>may not</b> 434 * throw an exception in response to any of {@code elements} being null or of the wrong type. 435 */ 436 @CanIgnoreReturnValue 437 @Override 438 boolean removeAll(Collection<?> c); 439 440 /** 441 * {@inheritDoc} 442 * 443 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 444 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 445 * multiset for every occurrence in {@code c}, see {@link Multisets#retainOccurrences(Multiset, 446 * Multiset)}. 447 * 448 * <p>This method refines {@link Collection#retainAll} to further specify that it <b>may not</b> 449 * throw an exception in response to any of {@code elements} being null or of the wrong type. 450 * 451 * @see Multisets#retainOccurrences(Multiset, Multiset) 452 */ 453 @CanIgnoreReturnValue 454 @Override 455 boolean retainAll(Collection<?> c); 456 457 /** 458 * {@inheritDoc} 459 * 460 * <p>Elements that occur multiple times in the multiset will be passed to the {@code Consumer} 461 * correspondingly many times, though not necessarily sequentially. 462 */ 463 @Override 464 default void forEach(Consumer<? super E> action) { 465 checkNotNull(action); 466 entrySet() 467 .forEach( 468 entry -> { 469 E elem = entry.getElement(); 470 int count = entry.getCount(); 471 for (int i = 0; i < count; i++) { 472 action.accept(elem); 473 } 474 }); 475 } 476 477 @Override 478 default Spliterator<E> spliterator() { 479 return Multisets.spliteratorImpl(this); 480 } 481}