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