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.GwtCompatible; 022import com.google.errorprone.annotations.CanIgnoreReturnValue; 023import com.google.errorprone.annotations.CompatibleWith; 024import java.util.Collection; 025import java.util.Collections; 026import java.util.Iterator; 027import java.util.List; 028import java.util.Set; 029import java.util.Spliterator; 030import java.util.function.Consumer; 031import java.util.function.ObjIntConsumer; 032import org.jspecify.annotations.Nullable; 033 034/** 035 * A collection that supports order-independent equality, like {@link Set}, but may have duplicate 036 * elements. A multiset is also sometimes called a <i>bag</i>. 037 * 038 * <p>Elements of a multiset that are equal to one another are referred to as <i>occurrences</i> of 039 * the same single element. The total number of occurrences of an element in a multiset is called 040 * the <i>count</i> of that element (the terms "frequency" and "multiplicity" are equivalent, but 041 * not used in this API). Since the count of an element is represented as an {@code int}, a multiset 042 * may never contain more than {@link Integer#MAX_VALUE} occurrences of any one element. 043 * 044 * <p>{@code Multiset} refines the specifications of several methods from {@code Collection}. It 045 * also defines an additional query operation, {@link #count}, which returns the count of an 046 * element. There are five new bulk-modification operations, for example {@link #add(Object, int)}, 047 * to add or remove multiple occurrences of an element at once, or to set the count of an element to 048 * a specific value. These modification operations are optional, but implementations which support 049 * the standard collection operations {@link #add(Object)} or {@link #remove(Object)} are encouraged 050 * to implement the related methods as well. Finally, two collection views are provided: {@link 051 * #elementSet} contains the distinct elements of the multiset "with duplicates collapsed", and 052 * {@link #entrySet} is similar but contains {@link Entry Multiset.Entry} instances, each providing 053 * both a distinct element and the count of that element. 054 * 055 * <p>In addition to these required methods, implementations of {@code Multiset} are expected to 056 * provide two {@code static} creation methods: {@code create()}, returning an empty multiset, and 057 * {@code create(Iterable<? extends E>)}, returning a multiset containing the given initial 058 * elements. This is simply a refinement of {@code Collection}'s constructor recommendations. 059 * 060 * <p>As with other collection types, the modification operations are optional, and should throw 061 * {@link UnsupportedOperationException} when they are not implemented. Most implementations should 062 * support either all add operations or none of them, all removal operations or none of them, and if 063 * and only if all of these are supported, the {@code setCount} methods as well. 064 * 065 * <p>A multiset uses {@link Object#equals} to determine whether two instances should be considered 066 * "the same," <i>unless specified otherwise</i> by the implementation. 067 * 068 * <p><b>Warning:</b> as with normal {@link Set}s, it is almost always a bad idea to modify an 069 * element (in a way that affects its {@link Object#equals} behavior) while it is contained in a 070 * multiset. Undefined behavior and bugs will result. 071 * 072 * <h3>Implementations</h3> 073 * 074 * <ul> 075 * <li>{@link ImmutableMultiset} 076 * <li>{@link ImmutableSortedMultiset} 077 * <li>{@link HashMultiset} 078 * <li>{@link LinkedHashMultiset} 079 * <li>{@link TreeMultiset} 080 * <li>{@link EnumMultiset} 081 * <li>{@link ConcurrentHashMultiset} 082 * </ul> 083 * 084 * <p>If your values may be zero, negative, or outside the range of an int, you may wish to use 085 * {@link com.google.common.util.concurrent.AtomicLongMap} instead. Note, however, that unlike 086 * {@code Multiset}, {@code AtomicLongMap} does not automatically remove zeros. 087 * 088 * <p>See the Guava User Guide article on <a href= 089 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#multiset">{@code Multiset}</a>. 090 * 091 * @author Kevin Bourrillion 092 * @since 2.0 093 */ 094@GwtCompatible 095public interface Multiset<E extends @Nullable Object> extends Collection<E> { 096 // Query Operations 097 098 /** 099 * Returns the total number of all occurrences of all elements in this multiset. 100 * 101 * <p><b>Note:</b> this method does not return the number of <i>distinct elements</i> in the 102 * multiset, which is given by {@code entrySet().size()}. 103 */ 104 @Override 105 int size(); 106 107 /** 108 * Returns the number of occurrences of an element in this multiset (the <i>count</i> of the 109 * element). Note that for an {@link Object#equals}-based multiset, this gives the same result as 110 * {@link Collections#frequency} (which would presumably perform more poorly). 111 * 112 * <p><b>Note:</b> the utility method {@link Iterables#frequency} generalizes this operation; it 113 * correctly delegates to this method when dealing with a multiset, but it can also accept any 114 * other iterable type. 115 * 116 * @param element the element to count occurrences of 117 * @return the number of occurrences of the element in this multiset; possibly zero but never 118 * negative 119 */ 120 int count(@CompatibleWith("E") @Nullable Object element); 121 122 // Bulk Operations 123 124 /** 125 * Adds a number of occurrences of an element to this multiset. Note that if {@code occurrences == 126 * 1}, this method has the identical effect to {@link #add(Object)}. This method is functionally 127 * equivalent (except in the case of overflow) to the call {@code 128 * addAll(Collections.nCopies(element, occurrences))}, which would presumably perform much more 129 * poorly. 130 * 131 * @param element the element to add occurrences of; may be null only if explicitly allowed by the 132 * implementation 133 * @param occurrences the number of occurrences of the element to add. May be zero, in which case 134 * no change will be made. 135 * @return the count of the element before the operation; possibly zero 136 * @throws IllegalArgumentException if {@code occurrences} is negative, or if this operation would 137 * result in more than {@link Integer#MAX_VALUE} occurrences of the element 138 * @throws NullPointerException if {@code element} is null and this implementation does not permit 139 * null elements. Note that if {@code occurrences} is zero, the implementation may opt to 140 * return normally. 141 */ 142 @CanIgnoreReturnValue 143 int add(@ParametricNullness E element, int occurrences); 144 145 /** 146 * Adds a single occurrence of the specified element to this multiset. 147 * 148 * <p>This method refines {@link Collection#add}, which only <i>ensures</i> the presence of the 149 * element, to further specify that a successful call must always increment the count of the 150 * element, and the overall size of the collection, by one. 151 * 152 * <p>To both add the element and obtain the previous count of that element, use {@link 153 * #add(Object, int) add}{@code (element, 1)} instead. 154 * 155 * @param element the element to add one occurrence of; may be null only if explicitly allowed by 156 * the implementation 157 * @return {@code true} always, since this call is required to modify the multiset, unlike other 158 * {@link Collection} types 159 * @throws NullPointerException if {@code element} is null and this implementation does not permit 160 * null elements 161 * @throws IllegalArgumentException if {@link Integer#MAX_VALUE} occurrences of {@code element} 162 * are already contained in this multiset 163 */ 164 @CanIgnoreReturnValue 165 @Override 166 boolean add(@ParametricNullness E element); 167 168 /** 169 * Removes a number of occurrences of the specified element from this multiset. If the multiset 170 * contains fewer than this number of occurrences to begin with, all occurrences will be removed. 171 * Note that if {@code occurrences == 1}, this is functionally equivalent to the call {@code 172 * remove(element)}. 173 * 174 * @param element the element to conditionally remove occurrences of 175 * @param occurrences the number of occurrences of the element to remove. May be zero, in which 176 * case no change will be made. 177 * @return the count of the element before the operation; possibly zero 178 * @throws IllegalArgumentException if {@code occurrences} is negative 179 */ 180 @CanIgnoreReturnValue 181 int remove(@CompatibleWith("E") @Nullable Object element, int occurrences); 182 183 /** 184 * Removes a <i>single</i> occurrence of the specified element from this multiset, if present. 185 * 186 * <p>This method refines {@link Collection#remove} to further specify that it <b>may not</b> 187 * throw an exception in response to {@code element} being null or of the wrong type. 188 * 189 * <p>To both remove the element and obtain the previous count of that element, use {@link 190 * #remove(Object, int) remove}{@code (element, 1)} instead. 191 * 192 * @param element the element to remove one occurrence of 193 * @return {@code true} if an occurrence was found and removed 194 */ 195 @CanIgnoreReturnValue 196 @Override 197 boolean remove(@Nullable Object element); 198 199 /** 200 * Adds or removes the necessary occurrences of an element such that the element attains the 201 * desired count. 202 * 203 * @param element the element to add or remove occurrences of; may be null only if explicitly 204 * allowed by the implementation 205 * @param count the desired count of the element in this multiset 206 * @return the count of the element before the operation; possibly zero 207 * @throws IllegalArgumentException if {@code count} is negative 208 * @throws NullPointerException if {@code element} is null and this implementation does not permit 209 * null elements. Note that if {@code count} is zero, the implementor may optionally return 210 * zero instead. 211 */ 212 @CanIgnoreReturnValue 213 int setCount(@ParametricNullness E element, int count); 214 215 /** 216 * Conditionally sets the count of an element to a new value, as described in {@link 217 * #setCount(Object, int)}, provided that the element has the expected current count. If the 218 * current count is not {@code oldCount}, no change is made. 219 * 220 * @param element the element to conditionally set the count of; may be null only if explicitly 221 * allowed by the implementation 222 * @param oldCount the expected present count of the element in this multiset 223 * @param newCount the desired count of the element in this multiset 224 * @return {@code true} if the condition for modification was met. This implies that the multiset 225 * was indeed modified, unless {@code oldCount == newCount}. 226 * @throws IllegalArgumentException if {@code oldCount} or {@code newCount} is negative 227 * @throws NullPointerException if {@code element} is null and the implementation does not permit 228 * null elements. Note that if {@code oldCount} and {@code newCount} are both zero, the 229 * implementor may optionally return {@code true} instead. 230 */ 231 @CanIgnoreReturnValue 232 boolean setCount(@ParametricNullness E element, int oldCount, int newCount); 233 234 // Views 235 236 /** 237 * Returns the set of distinct elements contained in this multiset. The element set is backed by 238 * the same data as the multiset, so any change to either is immediately reflected in the other. 239 * The order of the elements in the element set is unspecified. 240 * 241 * <p>If the element set supports any removal operations, these necessarily cause <b>all</b> 242 * occurrences of the removed element(s) to be removed from the multiset. Implementations are not 243 * expected to support the add operations, although this is possible. 244 * 245 * <p>A common use for the element set is to find the number of distinct elements in the multiset: 246 * {@code elementSet().size()}. 247 * 248 * @return a view of the set of distinct elements in this multiset 249 */ 250 Set<E> elementSet(); 251 252 /** 253 * Returns a view of the contents of this multiset, grouped into {@code Multiset.Entry} instances, 254 * each providing an element of the multiset and the count of that element. This set contains 255 * exactly one entry for each distinct element in the multiset (thus it always has the same size 256 * as the {@link #elementSet}). The order of the elements in the element set is unspecified. 257 * 258 * <p>The entry set is backed by the same data as the multiset, so any change to either is 259 * immediately reflected in the other. However, multiset changes may or may not be reflected in 260 * any {@code Entry} instances already retrieved from the entry set (this is 261 * implementation-dependent). Furthermore, implementations are not required to support 262 * modifications to the entry set at all, and the {@code Entry} instances themselves don't even 263 * have methods for modification. See the specific implementation class for more details on how 264 * its entry set handles modifications. 265 * 266 * @return a set of entries representing the data of this multiset 267 */ 268 Set<Entry<E>> entrySet(); 269 270 /** 271 * An unmodifiable element-count pair for a multiset. The {@link Multiset#entrySet} method returns 272 * a view of the multiset whose elements are of this class. A multiset implementation may return 273 * Entry instances that are either live "read-through" views to the Multiset, or immutable 274 * snapshots. Note that this type is unrelated to the similarly-named type {@code Map.Entry}. 275 * 276 * @since 2.0 277 */ 278 interface Entry<E extends @Nullable Object> { 279 280 /** 281 * Returns the multiset element corresponding to this entry. Multiple calls to this method 282 * always return the same instance. 283 * 284 * @return the element corresponding to this entry 285 */ 286 @ParametricNullness 287 E getElement(); 288 289 /** 290 * Returns the count of the associated element in the underlying multiset. This count may either 291 * be an unchanging snapshot of the count at the time the entry was retrieved, or a live view of 292 * the current count of the element in the multiset, depending on the implementation. Note that 293 * in the former case, this method can never return zero, while in the latter, it will return 294 * zero if all occurrences of the element were since removed from the multiset. 295 * 296 * @return the count of the element; never negative 297 */ 298 int getCount(); 299 300 /** 301 * {@inheritDoc} 302 * 303 * <p>Returns {@code true} if the given object is also a multiset entry and the two entries 304 * represent the same element and count. That is, two entries {@code a} and {@code b} are equal 305 * if: 306 * 307 * <pre>{@code 308 * Objects.equal(a.getElement(), b.getElement()) 309 * && a.getCount() == b.getCount() 310 * }</pre> 311 */ 312 @Override 313 // TODO(kevinb): check this wrt TreeMultiset? 314 boolean equals(@Nullable Object o); 315 316 /** 317 * {@inheritDoc} 318 * 319 * <p>The hash code of a multiset entry for element {@code element} and count {@code count} is 320 * defined as: 321 * 322 * <pre>{@code 323 * ((element == null) ? 0 : element.hashCode()) ^ count 324 * }</pre> 325 */ 326 @Override 327 int hashCode(); 328 329 /** 330 * Returns the canonical string representation of this entry, defined as follows. If the count 331 * for this entry is one, this is simply the string representation of the corresponding element. 332 * Otherwise, it is the string representation of the element, followed by the three characters 333 * {@code " x "} (space, letter x, space), followed by the count. 334 */ 335 @Override 336 String toString(); 337 } 338 339 /** 340 * Runs the specified action for each distinct element in this multiset, and the number of 341 * occurrences of that element. For some {@code Multiset} implementations, this may be more 342 * efficient than iterating over the {@link #entrySet()} either explicitly or with {@code 343 * entrySet().forEach(action)}. 344 * 345 * @since 21.0 346 */ 347 default void forEachEntry(ObjIntConsumer<? super E> action) { 348 checkNotNull(action); 349 entrySet().forEach(entry -> action.accept(entry.getElement(), entry.getCount())); 350 } 351 352 // Comparison and hashing 353 354 /** 355 * Compares the specified object with this multiset for equality. Returns {@code true} if the 356 * given object is also a multiset and contains equal elements with equal counts, regardless of 357 * order. 358 */ 359 @Override 360 // TODO(kevinb): caveats about equivalence-relation? 361 boolean equals(@Nullable Object object); 362 363 /** 364 * Returns the hash code for this multiset. This is defined as the sum of 365 * 366 * <pre>{@code 367 * ((element == null) ? 0 : element.hashCode()) ^ count(element) 368 * }</pre> 369 * 370 * <p>over all distinct elements in the multiset. It follows that a multiset and its entry set 371 * always have the same hash code. 372 */ 373 @Override 374 int hashCode(); 375 376 /** 377 * {@inheritDoc} 378 * 379 * <p>It is recommended, though not mandatory, that this method return the result of invoking 380 * {@link #toString} on the {@link #entrySet}, yielding a result such as {@code [a x 3, c, d x 2, 381 * e]}. 382 */ 383 @Override 384 String toString(); 385 386 // Refined Collection Methods 387 388 /** 389 * {@inheritDoc} 390 * 391 * <p>Elements that occur multiple times in the multiset will appear multiple times in this 392 * iterator, though not necessarily sequentially. 393 */ 394 @Override 395 Iterator<E> iterator(); 396 397 /** 398 * Determines whether this multiset contains the specified element. 399 * 400 * <p>This method refines {@link Collection#contains} to further specify that it <b>may not</b> 401 * throw an exception in response to {@code element} being null or of the wrong type. 402 * 403 * @param element the element to check for 404 * @return {@code true} if this multiset contains at least one occurrence of the element 405 */ 406 @Override 407 boolean contains(@Nullable Object element); 408 409 /** 410 * Returns {@code true} if this multiset contains at least one occurrence of each element in the 411 * specified collection. 412 * 413 * <p>This method refines {@link Collection#containsAll} 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 * <p><b>Note:</b> this method does not take into account the occurrence count of an element in 417 * the two collections; it may still return {@code true} even if {@code elements} contains several 418 * occurrences of an element and this multiset contains only one. This is no different than any 419 * other collection type like {@link List}, but it may be unexpected to the user of a multiset. 420 * 421 * @param elements the collection of elements to be checked for containment in this multiset 422 * @return {@code true} if this multiset contains at least one occurrence of each element 423 * contained in {@code elements} 424 * @throws NullPointerException if {@code elements} is null 425 */ 426 @Override 427 boolean containsAll(Collection<?> elements); 428 429 /** 430 * {@inheritDoc} 431 * 432 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 433 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 434 * multiset for every occurrence in {@code c}, see {@link Multisets#removeOccurrences(Multiset, 435 * Multiset)}. 436 * 437 * <p>This method refines {@link Collection#removeAll} to further specify that it <b>may not</b> 438 * throw an exception in response to any of {@code elements} being null or of the wrong type. 439 */ 440 @CanIgnoreReturnValue 441 @Override 442 boolean removeAll(Collection<?> c); 443 444 /** 445 * {@inheritDoc} 446 * 447 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 448 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 449 * multiset for every occurrence in {@code c}, see {@link Multisets#retainOccurrences(Multiset, 450 * Multiset)}. 451 * 452 * <p>This method refines {@link Collection#retainAll} to further specify that it <b>may not</b> 453 * throw an exception in response to any of {@code elements} being null or of the wrong type. 454 * 455 * @see Multisets#retainOccurrences(Multiset, Multiset) 456 */ 457 @CanIgnoreReturnValue 458 @Override 459 boolean retainAll(Collection<?> c); 460 461 /** 462 * {@inheritDoc} 463 * 464 * <p>Elements that occur multiple times in the multiset will be passed to the {@code Consumer} 465 * correspondingly many times, though not necessarily sequentially. 466 */ 467 @Override 468 default void forEach(Consumer<? super E> action) { 469 checkNotNull(action); 470 entrySet() 471 .forEach( 472 entry -> { 473 E elem = entry.getElement(); 474 int count = entry.getCount(); 475 for (int i = 0; i < count; i++) { 476 action.accept(elem); 477 } 478 }); 479 } 480 481 @Override 482 default Spliterator<E> spliterator() { 483 return Multisets.spliteratorImpl(this); 484 } 485}