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 javax.annotation.CheckForNull; 033import org.checkerframework.checker.nullness.qual.Nullable; 034 035/** 036 * A collection that supports order-independent equality, like {@link Set}, but may have duplicate 037 * elements. A multiset is also sometimes called a <i>bag</i>. 038 * 039 * <p>Elements of a multiset that are equal to one another are referred to as <i>occurrences</i> of 040 * the same single element. The total number of occurrences of an element in a multiset is called 041 * the <i>count</i> of that element (the terms "frequency" and "multiplicity" are equivalent, but 042 * not used in this API). Since the count of an element is represented as an {@code int}, a multiset 043 * may never contain more than {@link Integer#MAX_VALUE} occurrences of any one element. 044 * 045 * <p>{@code Multiset} refines the specifications of several methods from {@code Collection}. It 046 * also defines an additional query operation, {@link #count}, which returns the count of an 047 * element. There are five new bulk-modification operations, for example {@link #add(Object, int)}, 048 * to add or remove multiple occurrences of an element at once, or to set the count of an element to 049 * a specific value. These modification operations are optional, but implementations which support 050 * the standard collection operations {@link #add(Object)} or {@link #remove(Object)} are encouraged 051 * to implement the related methods as well. Finally, two collection views are provided: {@link 052 * #elementSet} contains the distinct elements of the multiset "with duplicates collapsed", and 053 * {@link #entrySet} is similar but contains {@link Entry Multiset.Entry} instances, each providing 054 * both a distinct element and the count of that element. 055 * 056 * <p>In addition to these required methods, implementations of {@code Multiset} are expected to 057 * provide two {@code static} creation methods: {@code create()}, returning an empty multiset, and 058 * {@code create(Iterable<? extends E>)}, returning a multiset containing the given initial 059 * elements. This is simply a refinement of {@code Collection}'s constructor recommendations. 060 * 061 * <p>As with other collection types, the modification operations are optional, and should throw 062 * {@link UnsupportedOperationException} when they are not implemented. Most implementations should 063 * support either all add operations or none of them, all removal operations or none of them, and if 064 * and only if all of these are supported, the {@code setCount} methods as well. 065 * 066 * <p>A multiset uses {@link Object#equals} to determine whether two instances should be considered 067 * "the same," <i>unless specified otherwise</i> by the implementation. 068 * 069 * <p><b>Warning:</b> as with normal {@link Set}s, it is almost always a bad idea to modify an 070 * element (in a way that affects its {@link Object#equals} behavior) while it is contained in a 071 * multiset. Undefined behavior and bugs will result. 072 * 073 * <h3>Implementations</h3> 074 * 075 * <ul> 076 * <li>{@link ImmutableMultiset} 077 * <li>{@link ImmutableSortedMultiset} 078 * <li>{@link HashMultiset} 079 * <li>{@link LinkedHashMultiset} 080 * <li>{@link TreeMultiset} 081 * <li>{@link EnumMultiset} 082 * <li>{@link ConcurrentHashMultiset} 083 * </ul> 084 * 085 * <p>If your values may be zero, negative, or outside the range of an int, you may wish to use 086 * {@link com.google.common.util.concurrent.AtomicLongMap} instead. Note, however, that unlike 087 * {@code Multiset}, {@code AtomicLongMap} does not automatically remove zeros. 088 * 089 * <p>See the Guava User Guide article on <a href= 090 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#multiset">{@code Multiset}</a>. 091 * 092 * @author Kevin Bourrillion 093 * @since 2.0 094 */ 095@GwtCompatible 096public interface Multiset<E extends @Nullable Object> extends Collection<E> { 097 // Query Operations 098 099 /** 100 * Returns the total number of all occurrences of all elements in this multiset. 101 * 102 * <p><b>Note:</b> this method does not return the number of <i>distinct elements</i> in the 103 * multiset, which is given by {@code entrySet().size()}. 104 */ 105 @Override 106 int size(); 107 108 /** 109 * Returns the number of occurrences of an element in this multiset (the <i>count</i> of the 110 * element). Note that for an {@link Object#equals}-based multiset, this gives the same result as 111 * {@link Collections#frequency} (which would presumably perform more poorly). 112 * 113 * <p><b>Note:</b> the utility method {@link Iterables#frequency} generalizes this operation; it 114 * correctly delegates to this method when dealing with a multiset, but it can also accept any 115 * other iterable type. 116 * 117 * @param element the element to count occurrences of 118 * @return the number of occurrences of the element in this multiset; possibly zero but never 119 * negative 120 */ 121 int count(@CompatibleWith("E") @CheckForNull Object element); 122 123 // Bulk Operations 124 125 /** 126 * Adds a number of occurrences of an element to this multiset. Note that if {@code occurrences == 127 * 1}, this method has the identical effect to {@link #add(Object)}. This method is functionally 128 * equivalent (except in the case of overflow) to the call {@code 129 * addAll(Collections.nCopies(element, occurrences))}, which would presumably perform much more 130 * poorly. 131 * 132 * @param element the element to add occurrences of; may be null only if explicitly allowed by the 133 * implementation 134 * @param occurrences the number of occurrences of the element to add. May be zero, in which case 135 * no change will be made. 136 * @return the count of the element before the operation; possibly zero 137 * @throws IllegalArgumentException if {@code occurrences} is negative, or if this operation would 138 * result in more than {@link Integer#MAX_VALUE} occurrences of the element 139 * @throws NullPointerException if {@code element} is null and this implementation does not permit 140 * null elements. Note that if {@code occurrences} is zero, the implementation may opt to 141 * return normally. 142 */ 143 @CanIgnoreReturnValue 144 int add(@ParametricNullness E element, int occurrences); 145 146 /** 147 * Adds a single occurrence of the specified element to this multiset. 148 * 149 * <p>This method refines {@link Collection#add}, which only <i>ensures</i> the presence of the 150 * element, to further specify that a successful call must always increment the count of the 151 * element, and the overall size of the collection, by one. 152 * 153 * <p>To both add the element and obtain the previous count of that element, use {@link 154 * #add(Object, int) add}{@code (element, 1)} instead. 155 * 156 * @param element the element to add one occurrence of; may be null only if explicitly allowed by 157 * the implementation 158 * @return {@code true} always, since this call is required to modify the multiset, unlike other 159 * {@link Collection} types 160 * @throws NullPointerException if {@code element} is null and this implementation does not permit 161 * null elements 162 * @throws IllegalArgumentException if {@link Integer#MAX_VALUE} occurrences of {@code element} 163 * are already contained in this multiset 164 */ 165 @CanIgnoreReturnValue 166 @Override 167 boolean add(@ParametricNullness E element); 168 169 /** 170 * Removes a number of occurrences of the specified element from this multiset. If the multiset 171 * contains fewer than this number of occurrences to begin with, all occurrences will be removed. 172 * Note that if {@code occurrences == 1}, this is functionally equivalent to the call {@code 173 * remove(element)}. 174 * 175 * @param element the element to conditionally remove occurrences of 176 * @param occurrences the number of occurrences of the element to remove. May be zero, in which 177 * case no change will be made. 178 * @return the count of the element before the operation; possibly zero 179 * @throws IllegalArgumentException if {@code occurrences} is negative 180 */ 181 @CanIgnoreReturnValue 182 int remove(@CompatibleWith("E") @CheckForNull Object element, int occurrences); 183 184 /** 185 * Removes a <i>single</i> occurrence of the specified element from this multiset, if present. 186 * 187 * <p>This method refines {@link Collection#remove} to further specify that it <b>may not</b> 188 * throw an exception in response to {@code element} being null or of the wrong type. 189 * 190 * <p>To both remove the element and obtain the previous count of that element, use {@link 191 * #remove(Object, int) remove}{@code (element, 1)} instead. 192 * 193 * @param element the element to remove one occurrence of 194 * @return {@code true} if an occurrence was found and removed 195 */ 196 @CanIgnoreReturnValue 197 @Override 198 boolean remove(@CheckForNull Object element); 199 200 /** 201 * Adds or removes the necessary occurrences of an element such that the element attains the 202 * desired count. 203 * 204 * @param element the element to add or remove occurrences of; may be null only if explicitly 205 * allowed by the implementation 206 * @param count the desired count of the element in this multiset 207 * @return the count of the element before the operation; possibly zero 208 * @throws IllegalArgumentException if {@code count} is negative 209 * @throws NullPointerException if {@code element} is null and this implementation does not permit 210 * null elements. Note that if {@code count} is zero, the implementor may optionally return 211 * zero instead. 212 */ 213 @CanIgnoreReturnValue 214 int setCount(@ParametricNullness E element, int count); 215 216 /** 217 * Conditionally sets the count of an element to a new value, as described in {@link 218 * #setCount(Object, int)}, provided that the element has the expected current count. If the 219 * current count is not {@code oldCount}, no change is made. 220 * 221 * @param element the element to conditionally set the count of; may be null only if explicitly 222 * allowed by the implementation 223 * @param oldCount the expected present count of the element in this multiset 224 * @param newCount the desired count of the element in this multiset 225 * @return {@code true} if the condition for modification was met. This implies that the multiset 226 * was indeed modified, unless {@code oldCount == newCount}. 227 * @throws IllegalArgumentException if {@code oldCount} or {@code newCount} is negative 228 * @throws NullPointerException if {@code element} is null and the implementation does not permit 229 * null elements. Note that if {@code oldCount} and {@code newCount} are both zero, the 230 * implementor may optionally return {@code true} instead. 231 */ 232 @CanIgnoreReturnValue 233 boolean setCount(@ParametricNullness E element, int oldCount, int newCount); 234 235 // Views 236 237 /** 238 * Returns the set of distinct elements contained in this multiset. The element set is backed by 239 * the same data as the multiset, so any change to either is immediately reflected in the other. 240 * The order of the elements in the element set is unspecified. 241 * 242 * <p>If the element set supports any removal operations, these necessarily cause <b>all</b> 243 * occurrences of the removed element(s) to be removed from the multiset. Implementations are not 244 * expected to support the add operations, although this is possible. 245 * 246 * <p>A common use for the element set is to find the number of distinct elements in the multiset: 247 * {@code elementSet().size()}. 248 * 249 * @return a view of the set of distinct elements in this multiset 250 */ 251 Set<E> elementSet(); 252 253 /** 254 * Returns a view of the contents of this multiset, grouped into {@code Multiset.Entry} instances, 255 * each providing an element of the multiset and the count of that element. This set contains 256 * exactly one entry for each distinct element in the multiset (thus it always has the same size 257 * as the {@link #elementSet}). The order of the elements in the element set is unspecified. 258 * 259 * <p>The entry set is backed by the same data as the multiset, so any change to either is 260 * immediately reflected in the other. However, multiset changes may or may not be reflected in 261 * any {@code Entry} instances already retrieved from the entry set (this is 262 * implementation-dependent). Furthermore, implementations are not required to support 263 * modifications to the entry set at all, and the {@code Entry} instances themselves don't even 264 * have methods for modification. See the specific implementation class for more details on how 265 * its entry set handles modifications. 266 * 267 * @return a set of entries representing the data of this multiset 268 */ 269 Set<Entry<E>> entrySet(); 270 271 /** 272 * An unmodifiable element-count pair for a multiset. The {@link Multiset#entrySet} method returns 273 * a view of the multiset whose elements are of this class. A multiset implementation may return 274 * Entry instances that are either live "read-through" views to the Multiset, or immutable 275 * snapshots. Note that this type is unrelated to the similarly-named type {@code Map.Entry}. 276 * 277 * @since 2.0 278 */ 279 interface Entry<E extends @Nullable Object> { 280 281 /** 282 * Returns the multiset element corresponding to this entry. Multiple calls to this method 283 * always return the same instance. 284 * 285 * @return the element corresponding to this entry 286 */ 287 @ParametricNullness 288 E getElement(); 289 290 /** 291 * Returns the count of the associated element in the underlying multiset. This count may either 292 * be an unchanging snapshot of the count at the time the entry was retrieved, or a live view of 293 * the current count of the element in the multiset, depending on the implementation. Note that 294 * in the former case, this method can never return zero, while in the latter, it will return 295 * zero if all occurrences of the element were since removed from the multiset. 296 * 297 * @return the count of the element; never negative 298 */ 299 int getCount(); 300 301 /** 302 * {@inheritDoc} 303 * 304 * <p>Returns {@code true} if the given object is also a multiset entry and the two entries 305 * represent the same element and count. That is, two entries {@code a} and {@code b} are equal 306 * if: 307 * 308 * <pre>{@code 309 * Objects.equal(a.getElement(), b.getElement()) 310 * && a.getCount() == b.getCount() 311 * }</pre> 312 */ 313 @Override 314 // TODO(kevinb): check this wrt TreeMultiset? 315 boolean equals(@CheckForNull Object o); 316 317 /** 318 * {@inheritDoc} 319 * 320 * <p>The hash code of a multiset entry for element {@code element} and count {@code count} is 321 * defined as: 322 * 323 * <pre>{@code 324 * ((element == null) ? 0 : element.hashCode()) ^ count 325 * }</pre> 326 */ 327 @Override 328 int hashCode(); 329 330 /** 331 * Returns the canonical string representation of this entry, defined as follows. If the count 332 * for this entry is one, this is simply the string representation of the corresponding element. 333 * Otherwise, it is the string representation of the element, followed by the three characters 334 * {@code " x "} (space, letter x, space), followed by the count. 335 */ 336 @Override 337 String toString(); 338 } 339 340 /** 341 * Runs the specified action for each distinct element in this multiset, and the number of 342 * occurrences of that element. For some {@code Multiset} implementations, this may be more 343 * efficient than iterating over the {@link #entrySet()} either explicitly or with {@code 344 * entrySet().forEach(action)}. 345 * 346 * @since 21.0 347 */ 348 default void forEachEntry(ObjIntConsumer<? super E> action) { 349 checkNotNull(action); 350 entrySet().forEach(entry -> action.accept(entry.getElement(), entry.getCount())); 351 } 352 353 // Comparison and hashing 354 355 /** 356 * Compares the specified object with this multiset for equality. Returns {@code true} if the 357 * given object is also a multiset and contains equal elements with equal counts, regardless of 358 * order. 359 */ 360 @Override 361 // TODO(kevinb): caveats about equivalence-relation? 362 boolean equals(@CheckForNull Object object); 363 364 /** 365 * Returns the hash code for this multiset. This is defined as the sum of 366 * 367 * <pre>{@code 368 * ((element == null) ? 0 : element.hashCode()) ^ count(element) 369 * }</pre> 370 * 371 * <p>over all distinct elements in the multiset. It follows that a multiset and its entry set 372 * always have the same hash code. 373 */ 374 @Override 375 int hashCode(); 376 377 /** 378 * {@inheritDoc} 379 * 380 * <p>It is recommended, though not mandatory, that this method return the result of invoking 381 * {@link #toString} on the {@link #entrySet}, yielding a result such as {@code [a x 3, c, d x 2, 382 * e]}. 383 */ 384 @Override 385 String toString(); 386 387 // Refined Collection Methods 388 389 /** 390 * {@inheritDoc} 391 * 392 * <p>Elements that occur multiple times in the multiset will appear multiple times in this 393 * iterator, though not necessarily sequentially. 394 */ 395 @Override 396 Iterator<E> iterator(); 397 398 /** 399 * Determines whether this multiset contains the specified element. 400 * 401 * <p>This method refines {@link Collection#contains} to further specify that it <b>may not</b> 402 * throw an exception in response to {@code element} being null or of the wrong type. 403 * 404 * @param element the element to check for 405 * @return {@code true} if this multiset contains at least one occurrence of the element 406 */ 407 @Override 408 boolean contains(@CheckForNull Object element); 409 410 /** 411 * Returns {@code true} if this multiset contains at least one occurrence of each element in the 412 * specified collection. 413 * 414 * <p>This method refines {@link Collection#containsAll} to further specify that it <b>may not</b> 415 * throw an exception in response to any of {@code elements} being null or of the wrong type. 416 * 417 * <p><b>Note:</b> this method does not take into account the occurrence count of an element in 418 * the two collections; it may still return {@code true} even if {@code elements} contains several 419 * occurrences of an element and this multiset contains only one. This is no different than any 420 * other collection type like {@link List}, but it may be unexpected to the user of a multiset. 421 * 422 * @param elements the collection of elements to be checked for containment in this multiset 423 * @return {@code true} if this multiset contains at least one occurrence of each element 424 * contained in {@code elements} 425 * @throws NullPointerException if {@code elements} is null 426 */ 427 @Override 428 boolean containsAll(Collection<?> elements); 429 430 /** 431 * {@inheritDoc} 432 * 433 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 434 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 435 * multiset for every occurrence in {@code c}, see {@link Multisets#removeOccurrences(Multiset, 436 * Multiset)}. 437 * 438 * <p>This method refines {@link Collection#removeAll} to further specify that it <b>may not</b> 439 * throw an exception in response to any of {@code elements} being null or of the wrong type. 440 */ 441 @CanIgnoreReturnValue 442 @Override 443 boolean removeAll(Collection<?> c); 444 445 /** 446 * {@inheritDoc} 447 * 448 * <p><b>Note:</b> This method ignores how often any element might appear in {@code c}, and only 449 * cares whether or not an element appears at all. If you wish to remove one occurrence in this 450 * multiset for every occurrence in {@code c}, see {@link Multisets#retainOccurrences(Multiset, 451 * Multiset)}. 452 * 453 * <p>This method refines {@link Collection#retainAll} to further specify that it <b>may not</b> 454 * throw an exception in response to any of {@code elements} being null or of the wrong type. 455 * 456 * @see Multisets#retainOccurrences(Multiset, Multiset) 457 */ 458 @CanIgnoreReturnValue 459 @Override 460 boolean retainAll(Collection<?> c); 461 462 /** 463 * {@inheritDoc} 464 * 465 * <p>Elements that occur multiple times in the multiset will be passed to the {@code Consumer} 466 * correspondingly many times, though not necessarily sequentially. 467 */ 468 @Override 469 default void forEach(Consumer<? super E> action) { 470 checkNotNull(action); 471 entrySet() 472 .forEach( 473 entry -> { 474 E elem = entry.getElement(); 475 int count = entry.getCount(); 476 for (int i = 0; i < count; i++) { 477 action.accept(elem); 478 } 479 }); 480 } 481 482 @Override 483 default Spliterator<E> spliterator() { 484 return Multisets.spliteratorImpl(this); 485 } 486}